PROJECT REPORT: MAXIMUM POWER POINT TRACKING OF PV ARRAYS IN

KUWAIT

Mohamed Ahmed Hassan El -Sayed Steven B. Leeb

Professor Professor of Electrical Engineering Kuwait University and
Computer Science, joint appointment in Mechanical
College of Engineering and Petroleum Engineering.
Electrical Engineering Department
MIT Laboratory for Electromagnetic
Tel : +965 2498-3347 and Electronics Systems. elsmah@hotmail.com
Tel: 6172539360 sbleeb@mit.edu

ABSTRACT

One technology to generate electricity from renewable resources is to use solar cells to
convert solar irradiance into electricity. Currently, photovoltaic (PV) energy generation has a
great commercial and academic interest. Recent studies indicate that in the medium to longer
term PV generation may become technically and commercially worldwide feasible to be
implemented for electricity generation. The main purpose of this report is to study the
performance of PV arrays under varying weather conditions in Kuwait. For this study a data
base of hourly solar radiation and temperature were collected for a period of six years. The
grid connected PV system converts sunlight directly into ac electricity to supply local loads
and inject the excess energy to the public grid. The main purpose of the system is to reduce
the electrical energy imported from the electric network. The dc output of the PV array is
converted into ac and injected into the grid through an interfacing inverter. The controller of
this system implements all the main regulation and protection functions such as:- Maximum
Power Point Tracking (MPPT), protection relay and detection of islanding operation.

New opportunities exist for deploying switched power electronic circuits with fresh thinking
about the role and use of energy storage components. During the course of this research, we
have explored the application of new power electronic approaches for cell level processing,
and for array configuration and MPPT tracking. This report presents some of the results from
this work.
Publications associated with this work:

1. Chang, Arthur, John J. Cooley, Steven B. Leeb, "Design Considerations in Distributed

MPPT Switched-Capacitor DC-DC Converters for PV Energy Extraction and Cost
Effectiveness," IEEE Applied Power Electronics Conference, Florida, February 2012

2. Chang, Arthur, S.B. Leeb, "Improved Transient Response Control Strategy and
Design Considerations for Switched-Capacitor (SC) Energy Buffer
Architectures," IEEE Applied Power Electronics Conference, California, March
2013
3. El-Sayed, M. and S. Leeb, “Evaluation of Maximum Power Point Tracking
Algorithms for Photovoltaic Electricity Generation in Kuwait,” International
Conference on Renewable Enegies and Power Quality, Cordoba, Spain, April 2014.
4. Gillman, M., J. Donnal, J. Paris, C. Schantz, U. Orji, S.B. Leeb, M. El-Sayed, K.
Wertz, S.Schertz, “Energy Accountability Using Non-Intrusive Load
Monitoring,” IEEE Sensors Journal, Vol. 14, No. 6, June 2014, pp.1923-1931.
 Table of Contents

Page
Chapter 1: Potential Utilization of Photovoltaic Electricity
1
Generation in Kuwait
1.1 Historical Background 1
1.2 Electrical Power System in Kuwait 3
1.2.1 Power Generation 3
1.2.1.1 Steam Turbine Units 4
1.2.1.2 Gas Turbine Units 4
1.3 Environmental considerations 9
1.4 Electrical Network 9
1.5 GCC Grid Systems Interconnection 14
1.6 Control Centers 16
1.7 Basic Requirements of the Supervisory Control Centers 18
1.8 MEW Future Projects 22
1.9 Renewable Energy in Kuwait 24
1.9.1 Solar Energy 27
1.9.1.1 Solar Energy Applications 27
1.9.1.2 Solar Energy Potential 27
1.9.2 Potential of Solar Power Generation in Kuwait 32
1.10 Renewable Power Forecasting Techniques 34
1.10.1 The Characteristics of Solar Radiation 35
1.11 Solar cooling 41

v
 1.11.1 Solar PV Air -Conditioning 43
1.11.2 Solar Thermal Air-conditioning 45
1.11.3 District cooling 48
1.12 Meteorological Data for Kuwait 51
1.12.1 Temperature 51
1.12.2 Cloud Coverage 52
1.12.2.1 Hourly Cloud Coverage 53
1.12.3 Precipitation 54
1.12.4 Humidity 54
1.12.5 Dew Point 55
1.12.6 Wind 57
1.12.7 Measurement of meteorological data 59
1.12.8 Global Solar Radiation 60
1.12.9 Wind Energy conversion 61
References 63
Chapter 2: Design and Modeling of Photovoltaic Systems 65
2.1 Definition 65
2.2 Photovoltaic Arrangements 65
2.2.1 Photovoltaic Cell 65
2.2.2 Photovoltaic Module 65
2.2.3 Photovoltaic Array 66
2.3 Materials Used in PV Cells 67
2.3.1 Single-Crystal Silicon 67
2.3.2 Polycrystalline Silicon 67

2.3.3 Gallium Arsenide (GaAs) 67

2.3.4 Cadmium Telluride (CdTe) 68

vi
 2.3.5 Copper Indium Diselenide (CuInSe2) 68
2.4 Characteristics of PV Cell 68
2.4.1 PV Module 71
2.4.2 PV Array 72
2.4.3 Effect of irradiance and temperature on the
72
module characteristics
2.5 Types of PV Systems 73
2.5.1 Stand-Alone Systems 73
2.5.2 Grid Connected PV Systems 75
2.6 Modeling of PV Systems 77
2.6.1 The Four-Parameter Single-Diode Equivalent
79
Circuit Model
2.6.1.1 Thermal Model of the PV Module 81
2.6.2 Simulink Implementation of the PV Model 82
2.6.3 Simulation of a PV module using the Simulink
85
Program
References 89
Error!
Chapter 3: Improvement of PV Panel Performance Using
Bookmark
Switched Capacitor Converters
not defined.
3.1 Improvement of Energy Extraction from Photovoltaic
Panels with Multilevel Output DC-DC Switched 90

vii
Capacitor Converters
3.1.1 Introduction 90
3.1.1.1 System Overview 92
3.1.1.2 Switched Capacitor Benefits 92

3.1.1.3 Total Efficiency 93

3.1.2 Maximum Power Point Tracking 94

3.1.2.1 PV Model 95
3.1.2.2 Local Maximum Power Point
96
Tracking
3.1.2.3 Global Maximum Power Point
98
Tracking
3.1.2.4 Statistical Performance Evaluation 98
3.1.2.5 Effect of spatial panel separation 100
3.1.2.6 Non-integral level selections 100
3.1.3 Switched Capacitor Implementation 101
3.1.3.1 Efficient Switching Patterns 101
3.1.3.2 Linear Modeling 102
3.1.3.3 Switching Speed Limit Definitions 103
3.1.3.4 Switching Loss 106
3.1.3.5 Inherent Features 107
3.1.3.6 Gate Drive 107
3.1.4 Design Example 109
3.1.4.1 Number of Levels 109
3.1.4.2 MOSFET Choice and Switching
109

viii
Frequency Optimization
3.1.4.3 Power Diode 111
3.1.4.4 Simulated Prototype 112
3.1.4.5 DC AC Dynamics 115
3.2 Improved Transient Response Control Strategy and
Design Considerations for Switched- Capacitor (SC) 117
Energy Buffer Architectures
3.2.1 Introduction 117
3.2.2 Control Algorithms 119
3.2.2.1 Bus-Voltage Monitoring, Finite State
120
Machine Control
3.2.2.2 Supporting Capacitor Monitoring,
122
Timing Interval Control
3.2.3 SC Energy Buffer Design Considerations
123

3.2.3.1 Capacitor Configurations 126

3.2.3.2 Switching Topology Tradeoffs 129
3.2.4 Two-Step Control Strategy 132
3.2.4.1 Capacitor Participation Optimization 133
3.2.4.2 Switch Timing Determination 134
3.2.4.3 Distortion and Phase Error 138
3.2.4.4 Pre-charge Circuit Requirement 139
3.2.4.5 Over- and Under-voltage Protection 140
3.2.5 Simulation Results 141
3.3 Conclusion 144

ix
References 145
Chapter 4: Maximum Power Point Tracking of PV Arrays in
149
Kuwait
4.1 Introduction 149
4.2 Maximum Power Point 150
4.3 Effect of Irradiance and Cell Temperature 151

4.4 Power Conditioning Devices 154

4.5 Power Electronics Devices for PV arrays 156
4.6 Boost Converter 157
4.6.1 Mode 1 of the Boost Converter Operation 158
4.6.2 Mode 2 of the Boost Converter Operation 158
4.7 DC-DC Converter Models 162
4.8 Maximum Power Point Tracking Algorithms 163
4.8.1 Constant Voltage Algorithm 165
4.8.2 Perturbation and Observation (P&O) Method 166
4.8.3 Incremental Conductance Method 167
4.8.4 Comparison of MPPT Algorithms 172
4.9 Simulation and Numerical Results 173
4.9.1 Dynamic response of the PV array for clear
176
day 28th June
4.9.2 Dynamic response of the PV array for
177
partially clouded day 19th Jan
4.9.3 Dynamic response of PV array for clouded
178

x
day 16th Nov
4.10 Conclusion 180
References 181

xi
 Chapter 1
Potential Utilization of Photovoltaic Electricity Generation in
Kuwait

This chapter presents electric energy resources in Kuwait and discusses the

development of Kuwait's electricity supplies. It also describes the electrical system in

Kuwait including generation, transmission and distribution subsystems. Moreover, the

available renewable energy such as solar energy and the associated radiation are

analyzed to justify the potential utilization of such clean resources.

1.1 Historical Background

State of Kuwait is located in the northwestern part of the Arabian Gulf, bordered by

Iraq from the north and Saudi Arabia from the south. It occupies 17,820 square

kilometers with a current population of a little more than 3 million. The population is

concentrated in and around Kuwait city along the gulf coast, in a radius of about 30

miles. The rest of the country is virtually unoccupied desert and oil fields [1.1].

The discovery of oil in the 1930's and then the first export of oil in 1946 have

transformed Kuwait into a modern country. Kuwait has an estimated oil reserve of 96

billion barrel, which accounts to 9% of the world's oil reserves. With the massive oil

revenues, the country was able to begin the civilization advancement in its

governmental utilities (electricity generation and water supply), education, health

care…etc. One of these utilities that will be considered in this report in details is the

electric utility. A Company called "The private electric company" first introduced

1
electricity in Kuwait in 1934; the company had a generating capacity of only 30 KW

at

200 Volts D.C. In 1949, the first 3 phase, 50 Hz AC power plant with 380/220 voltage

two generators, each with a capacity of 200 KW, was installed to replace the D.C.

system. In 1951, the Department of Electricity was established to be responsible for

electricity production. Later, the Department of Electricity was expanded as the

demand for electricity was increasing. That urged the establishment of what is now

known as the Ministry of Electricity and Water (MEW), which is a government

agency [1.2]. MEW had six dual purpose power plants before the Iraqi invasion in

1990. These plants had two functions to produce electricity and drinking water by

desalination process. The oldest dual power plant (Al-Shuwaikh plant) was

commissioned in 1954 with 4x7.5 MW capacity. The second oldest dual plant (Al-

Shuaiba plant) was commissioned between 1965 and 1969 with a total installed

capacity of 400 MW. These two plants were destroyed during the seven month Iraqi

occupation and not repaired due to unjustified repairing cost and their passed life time.

The other four power plants were repaired because they were relatively new and had

higher generation capacities than the previous older units. These four are Al-Shuaiba

South, Doha West, Doha East, and AlZour plants. These plants are fuel fired steam

plants using different types of fuel such as natural gas, heavy fuel oil, crude oil, and

gas oil. During peak load hours and emergency situations Gas turbines are also used,

in Al-Zour and Doha East plants. It is worth to mention that the electric demand is

increasing in an exponential rate and new Dual Purpose Plants (DPP) should be

installed to meet future needs. As a matter of fact, the DPP at Al-Sabiya is currently in

operation with an installed capacity of 2400 MW. The actual generation cost per kWh

2
reaches a value of 10 cent However, considering the high demand growth, it will be

necessary to build new power stations to meet the expected load. This requires

massive investment in electricity sector. In fact

Kuwait's oil export represents more than 90% of its total revenues. On the other side,

building new fuel-fired power plants considering the last time falling oil prices has a

great impact on the country's economy. This has prompted the government to utilize

non-conventional renewable energy for electricity generation and to encourage the

measures of energy conservation as well as rearranging the subsidiary policy of energy

consumptions.

1.2 Electrical Power System in Kuwait

1.2.1 Power Generation

The Electrical utility mainly employs Thermal Steam Turbines for the generation of

power needed to satisfy demand. However, Power Plants also include some Thermal

Gas Turbines that make up around 14.4% of total installed capacity. Otherwise, they

are kept as standby with a high degree of availability owing to gas turbines high

operational costs and low thermal efficiency. The Power Generation utility has over

the last five decades developed in quantity and quality. After the erection of the first

( 3X0.75 MW ) Steam Power Station in 1952, Power Plants capacities have increased

until they reached

2400 MW ( 8x300 MW) with the commissioning of Doha West Power Station in

1983/84 , Az-Zour South Power Station and Sabiya Power Station which is already

completed and in operation. The development of the peak load and the installed

capacity

3
are displayed in Table 1.1.
Table 1.1 Development of Peak demand and Installed Capacity in Kuwait
Year Peak Demand (MW) Installed Capacity (MW)
1978 1595 2128
1988 3920 7398
1998 5800 7389
2008 9710 11642
2010 10890 12145

1.2.1.1 Steam Turbine Units

These units comprise the large capacity units. Their capacities vary from 134 MW to

300 MW in all Power Stations. For example in year 2008, the total installed capacity

of these units is 8970 MW. The above units are operated according to the system

power demand. In general, the available and operational capacity will be maximum in

summer seasons as the electrical load demand increases with temperature rise and

minimum in the winter seasons, hence, routine annual maintenance of the above units

takes place during the winter season [1.2].

1.2.1.2 Gas Turbine Units

These are smaller capacity units ranging from 18 MW as in Doha East Power Station ,

42 MW in Shuwaikh Station, 28.2 MW in Doha West and 27.75 - 165 MW in Az-Zour

South Power Station and 41.7 MW as in Sabiya station. The total installed capacity is

2670.8 MW at high temperature operation. The above units are designed for normal

peak load operations with blackout start capability within 10 minutes, where it takes

more than five hours normally in case of steam turbine units. Table 1.2 (a) displays the

turbines capacities in the different power plants of Kuwait. The development of Power

4
Station’s Installed Capacity and the generated energy in the period from 1990-now is

displayed in Table 1.2 (b).

Table 1.2 (a) Turbine Capacity in the Power Plants

Power Plant Steam turbines capacities Gas turbines capacities (MW)
(MW)
Shuwaikh 6x42 MW= 252
Shuaiba South 6x 134 MW = 804
Doha East 7x 150 MW = 1050 6x18 MW = 108
Doha West 8x 300 MW = 2400 3x 28.2 MW = 84.6
Az-Zour South 8x 300 MW = 2400 4x27.5+8x130+5x165=1976
Subiya 8x 300 MW = 2400 6x41.7MW= 250.2

Table 1.2 (b) Development of Installed Capacity and Generated Energy in Kuwait
Power Energy Power Energy
Year Year
(MW) (MKWh) (MW) (MKWh)
1987 6696 18092 2000 9189 32323

1988 7398 19599 2001 9189 34299

1989 7411 21085 2002 9189 36362

1990 7283 18477 2003 9189 38577

1991 6898 10780 2004 9689 41257

1992 6898 16885 2005 10189 43734

1993 6898 20178 2006 10189 47605

1994 6898 22802 2007 10481 48754

1995 6898 23724 2008 11641 51749

1996 6898 25475 2010 12145 55810

1997 6898 26724

1998 7389 29984

1999 8289 31576

5
 Fig. 1.1 The energy consumed in different sectors in Kuwait

As indicated in Fig. 1.1 the domestic sector in private and investment building

represents 62% of the total energy consumption. Table 1.3 shows the maximum and

minimum daily consumption in kWh for each month of year 2008. The relative

humidity and maximum temperature as well as the percentage annual peak load

increase are given in Table 1.4. The following figures show the total load, temperature

and humidity at daily peak load for the summer months. The peak load reached 9710

MW on Saturday, 27th of July 2008 at 14:30, with a temperature of 50° C and relative

humidity of 5 % with an annual increase of 7.06 %. In addition, the variation of load

factor from year 1985-2000 is drawn in Fig. 1.2. It should be noted that the maximum

load on 15 July 2012 has a value of 11610 MW while the installed capacity reaches

about 13000 MW. The corresponding daily load curve at peak value is displayed in

Fig. 1.3. Moreover, the relative humidity and maximum temperature for the days of

peak load are displayed in Table 1.4 for the period from 1992-2008.

6
 Table 1.3 Maximum and Minimum Daily Consumption in kWh
Month Max. Consumption Date Min. Consumption Date
January 94027 16 Jan. 71684 04 Jan.
February 84712 04 Feb. 68660 29 Feb.
March 114211 04 Mar. 71047 07 Mar.
April 142315 29 Apr. 92530 04 Apr.
May 165995 29 May. 121973 02 May.
June 187150 29 Jun. 143253 13 Jun.
July 189070 27 Jul. 159034 11 Jul.
August 187856 03 Aug. 165679 19 Aug.
September 191182 08 Sep. 137652 30 Sep.
October 144193 12 Oct. 92933 11 Oct.
November 93296 01 Nov. 77736 07 Nay.
December 87591 2I Dec. 72169 09-0ec.

Table 1.4 Relative Humidity and Maximum Temperature at Peak Load

Max. Rel. Max. Temp. Percentage Max.
Humidity % at Peak Load Date of Annual Load Year
at Peak Load (°C) Increase (MW)
32 48 16.08.1992 3460 1992
28 48 17.08.1992
6 49 10.08.1993 19.1 4120 1993
8 45 01.06.1994 5.6 4350 1994
5 48 21.06.1995 8.7 4730 1995
28 43 02.09.1996 9.9 5200 1996
18 48 28.06.1997 3.1 5360 1997
10 49 19.08.1997 8.2 5800 1998
8 48 04.09.1999 6.2 6160 1999
5 50 28.08.2000 4.7 6450 2000
38 43 13.08.2001 4.7 6750 2001
7 50 22.07.2002 7.4 7250 2002
2 49 06.07.2003 3.2 7480 2003
16 50 26.07.2004 3.6 7750 2004
7 48 03.08.2004

7
7 51 17.07.2005 8.4 8400 2005
6 49 26.07.2006 6.0 8900 2006
6 47 03.09.2007 1.9 9070 2007
5 50 27.07.2008 9.1 9710 2008

Fig. 1.2 Load factor variation for the period from 1985-2008

Fig. 1.3 Peak load curve in 2012

8
1.3 Environmental considerations

It is very important that an Environmental Impact Assessment is carried-out by any

energy project planning. This includes site preparation, construction, operation,

maintenance and decommissioning phases. The general significant issues that needed

to be addressed include the following:

• Air pollution.

• Noise.

• Non-thermal water pollution.

• Thermal water pollution.

• Electromagnetic fields.

• Fuel storage areas and loading terminals.

• Waste management.

• Health and safety within and out-side the plant boundary.

• Ecosystem and its biotic and abiotic components.

Kuwait electrical load is characterized by high load in summer and low load in winter

according to the increase and decrease in the values of temperatures and relative

humidity.

1.4 Electrical Network

Currently Kuwait is covered by electric power networks system consisting of overhead

lines, underground cables and power transformer substations which ensure supply to

9
every consumer with electric energy including residential, industrial, commercial or

agricultural loads. The expansion of these networks has been carried out in parallel to

the rapid growth of the required electric demand. Furthermore, work is continuously

under way to strengthen and reinforce the different electrical networks to meet the

additional loads resulting from the change in consumption patterns.

A brief summary of the electrical networks expansion could be expressed as follows:

In 1934, electrical network started using + 200 V (Direct Current) O/H lines with

copper conductors on wooden poles. In 1949, the network voltage was changed to

380/220 V 50 Hz (A.C) after the expansion of the National Electricity Company

Services. Upon the transfer of this company to Government ownership in 1951 and the

erection of Shuwaikh power plant, the network voltage was raised from 380 V to 11

KV to extend the ranges of the network to satisfy the growing electrical load. Also 11

KV/415 V substations were constructed with the 11 KV underground cables and

overhead lines. As 11 KV network was unable to transmit the increasing power to

remote areas, a 33 KV transmission network was added in year 1958 comprising 33

KV O/H lines, underground cables and 33/11 KV substations in town and Salmiya.

Upon the erection of additional power plant in 1961-62, the 33 KV transmission

network was expanded with additional 33/11 KV sub-stations erected in Town,

Hawally and AlJahra. After the erection of Shuaiba North power plant there was a

need to raise the voltage to 132 KV. Consequently, the Ministry constructed the first

132 KV transmission line from Kuwait City to Shuaiba and consequently, the 5th Ring

Road

132/33 KV substation was constructed to transmit power to Shuaiba and Fahaheel.

10
Expansion and reinforcement of 132 KV network were continued in order to connect

Shuaiba South and then Doha East power plant. The 132 KV network served to

transmit power in bulk to the major 132/33 KV and 132/11 KV substations

constructed in the main load centers in the country [1.1].

With the continued growth of electrical demand and the corresponding increase in the

generating capacity, the 132 KV transmission network reached its economic and

technical limit. So it was decided to introduce a new 275 kV transmission network to

link the Power Stations with the 300/132 KV substations which will constitute major

supply centers in various areas. Fig. 1.4 shows the existing 275 kV network in year

2012.

1. E.H.V. Network:

A - Length of Underground 275 kV Cables - 216 KM.

B - Overhead 275 kV Lines - 712 KM.

C - Number of 275 kV Sub- Stations - 19.

2. H.V. Network:

A - Number of 132 KV Sub-Stations - 277.

B - Number of 33 KV Substations - 292.

C - Length of Underground 132 KV Cables - 2566 KM.

D - Length of Underground 33 KV Cables - 1531 KM.

E - Length of Overhead 132 KV Lines – 2162 KM.

F - Length of Overhead 33 KV Lines - 1827 KM.

3. M & L.V. Network 11 KV 415 V:

A - Number of Ring Main Substations - 7043.

11
 B - Number of Spur Substations - 5800.

C - Length of 11 KV Cables - 8784 KM.

D - Length of Low Tension Cables (415 V) – 22886 KM.

E - Length of Overhead 11 KV415 V and Low Tension Lines - 4415 KM.

Fig. 1.4 Layout of 275 kV network in Kuwait

4. Street Lighting:

A - Length of Streets with 30 M Masts - 168 KM.

B - Length of Streets with 22 M Masts – 263.7 KM.

C - Length of Streets with 12 M Masts and 16 M Masts – 1728.5 KM.

D - Streets with HPMV Lamps on 10 M Masts - 1833.2 KM.

E - Side Roads, Different Type of Lamps on 8 M Poles – 1887.1 KM.

F - Sub-Main Roads, Different Type of Lamps on 4M – 6M Poles – 1846.2

KM.

12
Most of the electrical consumption takes place at the L.V. level (415/240 volts) with

the exception of some industries where consumption takes place at medium voltage

(11

KV) or (3.3 KV). However, electric power generation takes place at the medium
voltage

(from 11 KV to 21 KV) and this voltage is not adequate to transmit large amounts of

power over long distances to the center of consumption. Hence, it is necessary to step

up the voltage at the Power Station to 132 or 275 kV in order to transmit the large

amounts of power to the main centers of consumption where the voltage is stepped

down to lower levels in suitable substations and the power distributed in smaller

amounts and in successive stage until the levels of power and voltage are suitable to

supply the consumers. Therefore, the electrical networks comprise the following types

of sub-stations:

• Step-up Substations at the Power Stations to raise the voltage from (11 - 21

KV) to 132 KV or 275 kV.

• Step-down Substations to lower the voltage from 275 kV to 132 KV.

• Step-down Substations to lower the voltage from 132 KV to 33 KV.

• Step-down Substations to lower the voltage from 132 KV to 11 KV.

• Step-down Substations to lower the voltage from 33 KV to 11 KV.

• Step-down Substations to lower the voltage from 11 KV to 415/240 Volts.

• These types of sub-stations are interconnected amongst themselves as well as

with the Power Stations on one side and with the consumers on the other side

13
 by various networks of different voltages levels. The following resume outlines

the main characteristics and functions of these networks:

• The primary transmission EHV network (275 kV) which consists mainly of

double circuit overhead lines having quadruple bundle conductors supported on

steel towers. In built-up areas, this network comprises 275 kV underground

cables of special design. This network has two functions - firstly to transmit

bulk

Power Station to major center of consumption (the 300/132 KV Substations) and

secondly, to interconnect the Power stations in perfect synchronism whereby all

generators maintain on identical speed.

• The secondary transmission HV Networks (132 KV and 33 KV) which consists

in open areas, of double circuit overhead lines having single of double-bundle

conductors supported on steel towers and in built-up areas of underground

cables of special design. The main function of this network is to transmit

medium amounts of power from the 300/132 KV and 132/33 KV substations to

the distribution network (11 KV and 415/240 Volts) and sometimes to large

consumers.

• Street-lighting network which consists of L.T. Cables emanating from 11 KV /

415 - 240 Volts sub-station to supply street lighting lanterns on poles and

masts.

14
1.5 GCC Grid Systems Interconnection

An idea to connect all six Gulf Cooperation Council Countries (GCCC), Kuwait,

Saudi Arabia, Qatar, Bahrain, United Arab Emirates and Oman, was put down in the

early eighties because Saudi Arabia uses the 60 Hz network, where the other five

countries uses the 50 HZ system. However, with the technological advancement of the

HVDC system, this idea gained further support in the nineties. Recognizing the

benefits of interconnection of their power grids, the six countries have agreed in 1997

to establish HVDC connection as economical and feasible alternative.

AC interconnection of the 50 Hz systems of Kuwait, Bahrain, Qatar, UAE and Oman

was built with a back-to-back HVDC interconnection to the 60 Hz Saudi Arabian

system. The interconnection was justified based on reserve sharing between the

systems and the opportunity for trading electricity between the member countries.

After

15
 Fig. 1.5 Route and layout of the GCC interconnection

investigating different options for financing the Project, it was decided to finance the

Project with funds from the member countries. The electrical grid system

interconnection between the GCC states is shown diagrammatically in Fig. 1.5. The

Interconnection Project consists of the following high voltage lines and the associated

substation:

• A double-circuit 400 kV, 50 Hz line from Al Zour (Kuwait) to Ghunan (Saudi

Arabia) with an intermediate connection at Fadhili (Saudi Arabia).

• A back-to-back HVDC interconnection to the Saudi Arabia 380 kV, 60 Hz,

system at Fadhili.

• A double circuit 400 kV comprising overhead lines and submarine link from

Ghunan to Al Jasra (Bahrain).

• A double circuit 400 kV line from Ghunan to Salwa (Saudi Arabia) and

associated substations.
• A double circuit 400 kV line from Salwa to Doha South (Qatar) and associated

substations.

• A Control Centre located at Ghunan.

• A double circuit 400 kV line from Salwa to Ghuwaifat (UAE).

• A double and a single circuit 220 kV line from Al Ouhah (UAE) to Al Wasset

(Oman).

A Control Center for the GCC Interconnection, which is also capable of

communication with the national centers of the Member Countries, will also be

installed and, when eventually completed, the interconnected substations will be

16
suitable for operation in a coordinated and stable manner through this Control Center.

The Center will provide the control and monitoring of the exchanged power between

the countries.

1.6 Control Centers

The fast development of the Electrical Networks in Kuwait and to co-ordinate the

activities of Power Generation, Transmission and Distribution and to secure the

continuity of the supply to the consumers in reliable and economic way, the Ministry

decided to construct National Supervisory Control Center and distribute Regional

Control Centers according to the geographical area and define their responsibilities as

described by:

1. National Control Center Jabriya:

This Center was constructed by Siemens in 1980 and upgraded by the same company
in

1994. This Center is responsible for the supervision and control of 275 kV, 132 kV

Electrical Network and supervises all Generating Stations in the country.

2. Jabriya District Control Center:

This Center started operating from 1972 and has been upgraded by Siemens on 1994.

This Center is responsible for the supervision and control of 132, 33 and 11 kV

Network which lies in the area surrounded by Fourth Ring Road in north and Gulf

Street on the east and Sixth Ring Road with Sabah Al Salem Area to the South and

AlGhazali Street with Riqaee and South Ardiya areas to the west.

3. Town District Control Center:

17
This Center started operating in 1988 and has been upgraded by Mitsubishi on
20/10/99.

This Center is responsible for the supervision and control of 132, 33, 11 kV Electrical

Network which lies in the area surrounded by Fourth Ring Road in the south and

AlGhazali Street to the west including Shuwaikh Industrial area and Hospitals area

and Gulf Street on the north and east.

4. Shuaiba District Control Center:

This Center was constructed by Siemens in 1988 and upgraded by the same company
on

30/10/99. This Center is responsible for the supervision and control of 132, 33, 11 kV

Electrical Network which lies in the area surrounded by Sixth Ring Road in the north

(excluding Sabah Al Salem Area) and the Kuwait Saudi border in the south and Gulf

Street on the east and Faisal Bin Abdul Aziz Motorway on the west.

5. Jahra District Control Center:

This Center constructed by Cogelex in 1994 and upgraded by the same company on

24/7/2002. This Center is responsible for the supervision and control of 132, 33, 11 kV

Electrical Network which covers all the area on western side starting from Sulaibikhat

and includes Ardiya, Andalus, and Sulaibiya.

1.7 Basic Requirements of the Supervisory Control Centers

The Control Centers require some very sensitive, accurate, reliable and advanced

systems which can be divided into the following categories:

• Information Collection Systems.

18
 • Telephony Systems and Peripherals.

• Communication System and Peripherals.

• Auxiliary Power Supply System and Peripherals.

• Computer System & Peripherals.

Next, a description of each category is given.

1. Information Collection System:

This system consists of a microprocessor with special programmes to collect and send

all required information and measurements from the Power Generating Stations and

Substations to the Control Centers and to receive the instructions issued from the

Control Centers and to ensure their correctness and follow up their execution. The

sending and receiving is to be very fast using different communication equipment.

2. Telephony System and Peripherals:

The MEW utilizes redundant communication facilities in different Power Generating

Stations and Substations to support the operating engineers for coordinating the

operation of the power system and taking the corrective measures in case of any

forced outage. Telephony network for each Center and are interconnected to each

other to increase the efficiency of these networks.

The Telephonic System of each Center consists of a main PABX installed in the

Control Center while the PAX’ s are usually installed in some of the major sub-

stations connected to the same Center. All PABX’s available in the control Centers are

connected to the Ministry of Communication Telephony network to have access to

direct outside communication in addition to the internal communication. It is to be

19
noted that the capacities of these PABX’s differ according to the operational

requirements of each Center. With regards to the Peripherals to the main PABX’s

consists of Multichannel Duplicated Voice Recording Equipment connected to the

telephonic lines and is used to record all the conversation exchanged between the

Control Room Operation Engineer and others round the clock. By this way the

operation engineer can analyze the instructions and to issue the required remedial

actions.

3. Communication System and Peripherals:

It has been considered in the designing of the Communication System that all major

substations data is transmitted via duplicate channels and the equipment differs

according to the routes used and the wired or wireless transmission media. Separate

Pilot Cables or communication cables are used with underground cables for power

transmission which are laid alongside on the same route for the transmission data in

the frequency range of 4 KHz and in case of Overhead Lines, these lines are used for

data transmission using Power Lines Carrier equipment which operate in the

frequency range of 35 – 500 KHz. With regards to wireless media, the equipment

installed is Microwave which operate in the frequency range of 7.5 GHz. Some major

substations collects the data of their satellites and transmits and receives the

information in the same way as described earlier. Digital Microwave Equipment

which operates in the frequency range 18, 19 Hz has been installed between one

substation and Shuaiba DCC. In case of substations located in far areas and for which

the previously described medias are not economical to be used then the equipment

using the Ultra High Frequency (UHF) wireless operating in the frequency range of

20
367.925 – 399.5 MHz is installed for the provision of telephonic services with the

Control Centers.

Ministry started in 1997 to use fiber optic technology for transmission of data and

telephony with a link between NCC and MEW Head office due to its superior features

such as the speed, capacity for data transmission and diversity of information transfer

(analog and digital) in addition to efficiency, integrity and availability of

communication channels.

Conventional ground wires, on eight 33 KV OHL links majoring 56 km have been

replaced with optical ground wires (OPGW). The replacement of conventional ground

wire for many existing 300 & 132 KV overhead lines with optical ground wires

(OPGW) is underway.

4. Auxiliary Power Supply System:

In view of the important role played by the supervisory control equipment and the

continuous operation even in case of loss of electrical supply to the station an auxiliary

power supply system should be provided in the Control Center and each station.

Auxiliary system in the substations consists of battery charger which converts the

electrical energy from the alternating current to the direct current and stores them in

battery. These batteries provide the required energy in case of loss of supply to the

charger for a period between 10 to 12 hours. The Auxiliary Power Supply System as

provided for the Control Centers consists of the following:

Uninterruptible Power Supply System which feeds the required power to the

computers and their peripherals.

21
Auxiliary Power Supply System for the Communication and Telephony Equipment

and their Peripherals which is identical to what is provided in the substations with the

only difference that their capacities are larger matching to the equipment requirements

present in the Control Centers.

5. Computer Systems and peripheral:

The duty of the Control centers depends mainly on the computer systems, which

communicate with the remote terminal units and power stations. These systems were

designed as Duplex System for work continuity. The National Control Center works

through special programs to display and analyze the data (SCADA Functions) in

addition to special programs for the Energy Management System. However the duty of

the district centers is restricted on data display and analysis (SCADA Functions)

besides the transmission of some data to the National Control Center in order to use it

for the Energy Management System.

6. Internet Service Provider:

The National Control Center is acting as the Internet Service Provider. A complete

domain has been registered for the use of the Ministry of Electricity & Water to serve

the e-mail services. An official web site has been developed under the name www.

mew.gov.kw. The site includes many general information such as Ministry of

Electricity & Water telephone directory, steps and procedure getting energy services,

water and electricity saving advises. Billing information, training programs and load

gauge.

7. Energy Saving Campaign:

22
National Control Center participated in the Energy Saving Campaign by exporting

information regarding energy consumption related to areas of Kuwait from the

National and District Control Centers SCADA system to Kuwait National Campaign

Control Room. This step contributed among others in reducing the consumptions of

the load.

8. Upgrade of Internet Infrastructure:

Currently, MEW is undertaking a project to upgrade the Infrastructure at the Ministry

and implementation of United Communication Systems in addition to enhance the

overall network security. The solution will combine upgrading the network

infrastructure at MEW head office by providing fully redundant, highly available and

secure network with 10 Gigabit Ethernet Fiber backbone. The solution will also

include installation of a pilot Unified Communication system with different types of

needed components and IP Phone. The complete system will be highly secure, voice

and video ready. IP Telephony and Video conferencing will be implanted as part of

Unified Communication.

1.8 MEW Future Projects

Government authorities are also planning to bring additional capacity over the next

few years, with the private sector expected to play a major role in this expansion.

According to the displayed data in section 2, peak demand for power is expected to

increase at a rate of 6-8% per annum over the next few years, reaching 14,256 MW by

2014. Peak demand usually occurs in the summer, when heavy use of cooling systems

can strain the power grid. According to some estimates, air conditioners account for

23
around 70% of Kuwait’s peak power demand. By the beginning of June all power

station will be committed according to cover the load. Table 1.5 depicts the future

power stations under study with the associated capacity to cover the future increasing

demand. Moreover, Table 1.6 indicates the total number of the required projects and

their associated costs.

Table 1.5 Future Expansion Planning in Power Plants

Project Capacity
Az-Zour Power Plant (4 Stages) 4800 MW+280 MIGD
Al-Julaia Power Planr 1000 MW
New Shuwaikh Power plant 2000 MW
New Shuaiba South Power Plant 1400 MW+50 MIGD
New Doha East Power Plant 2300+100 MIGD
Total to be added 11500 MW + 430 MIGD

Table 1.6 Expansion Planning Projects and the Required Cost

Total No. of Total Cost of
Sector
Projects Projects MKD

Planning & Training 9 85

Power Stations & Desalination plants 13 2800

Water Projects 27 1129

Control Centers & IT 20 158

Electrical Transmission 62 2400

Electrical Distribution 9 550

Operation & Maintenance Power Stations 8 325

Consumer Affairs 7 125

Technical Services & Main Workshops 6 35

24
 Total 134 7500
Indeed, in late June 2012 six gas-powered turbines have been activated, which will be

able to generate 1400 MW of electricity, equivalent to roughly 10% of Kuwait’s

current total capacity. The plant is expected to enter its full combined-cycle

commercial operations by the end of 2012, at which point its total capacity will be

around 2000 MW. A combined-cycle plant is more efficient because waste heat is

recovered and converted into steam, which is then used to drive a steam turbine to

produce additional power. Owned by the state, the facility will be operated and

maintained by General Electric for a period of seven years from the commencement of

full commercial operations.

The consumption of electrical energy per sector in Kuwait indicated that the domestic

sector has the highest share of this energy. In addition, there has been an escalating

increases in the demand for electrical energy and in the national cost for energy

production. Therefore, the government of Kuwait continues to encourage the

development of a code of practice for energy conservation in buildings and to utilize

renewable energy.

1.9 Renewable Energy in Kuwait

In its broadest sense, renewable energy can be understood as energy generated from

ecological resources which are naturally replenished. Renewable energy sources

include sunlight, wind, rain, tides and geothermal heat (for electricity generation),

organic-based biofuels (for heating and transportation) and traditional biomass (fuel

wood and agricultural waste) (for cooking and heating). The focus of this chapter will

be on renewable energy in the context of larger-scale electricity generation, which at a

25
GCC level has already achieved some success and has the potential to achieve

significantly more. Although nuclear power generation is often discussed alongside

renewable energy, it is technically not a renewable source of energy and accordingly

will not be discussed in this report other than to note that certain GCC countries have

announced an intention to include nuclear energy in their future energy mix.

Several renewable energy technologies have made significant penetrations in the

world energy market. These technologies include solar (solar thermal and

photovoltaic), wind, biomass, geothermal, hydro, wave, tidal and ocean thermal

resources applications. The increased interest and rapid growth in these applications

have been stimulated by a significant drop in cost over the previous decades, and

technical improvement in their efficiency, reliability and life time. Other advantages of

renewable energy applications are their modularity, large solar potential in developing

countries especially in MENA region, favorable land-use features and ease of

decommissioning.

The leading technical improvements in this field includes: manufacturing, innovations,

improvements in solar system design, conservation, efficiency improvements,

aerodynamic advances in wind turbine design and other mechanical and electrical

innovations. In this work, the renewable energy applications and its potentials in

MENA region is discussed and the needed steps to exploit and apply these sources in

Kuwait is studied.

The benefits of developing renewable energy in MENA region or in anywhere else can

be categorized in four headings: sustainability, environment, diversification, and

economics. The following are some of these benefits:

26
• Renewable energies are sustainable sources of energy. Renewable technologies

are designed to run on a virtually inexhaustible or replenishable supply of

natural fuels. The primary long term benefits of renewable technologies is that

once a renewable energy project has been constructed and fully operated, it

become a permanent component of the national energy system.

• Renewable sources promote energy diversification. Development of a diverse

portfolio of generation assess reduces a country dependence on any one

particular form of technology or fuel.

• Renewable energy technologies have a very small impact on environment

compared to fossil fuel. The discharge of unwanted or unhealthy substance in

air, ground and water commonly associated with other forms of energy use can

be reduced significantly by using renewable energies.

• Renewable energies have values beyond they generate. Renewable energy

systems are modular, flexible and can be installed anywhere and in any size.

Investment in locally available renewable energy generates more jobs, greater

earnings and higher output. The renewable energy industry provides a wide

range of employment opportunities, from high technology manufacturing of

PV components to maintenance jobs at solar thermal or wind systems. Several

renewable energy technologies and applications are of relevance to developing

countries. However, local and country-specific conditions including the

availability of renewable energy resources, cost of conventional energy, capital

cost of equipment, as well as other factors, influence the commercial viability

of renewable energy technologies. The following sections present brief

27
 description of renewable energy technologies that could be utilized in Kuwait

and its applications.

1.9.1 Solar Energy

1.9.1.1 Solar Energy Applications

Energy from the sun travels to the earth in the form of electromagnetic radiation

similar to radio waves, but in a different frequency range. In a clear day, the amount of

solar radiation available at earth surface in the direction of the sun is typically 1,000

W/m2. At any particular time, the available solar energy is primarily dependent upon

how high the sun in the sky and the current cloud condition. There are many ways that

solar energy can be used effectively. Applications of solar energy can be grouped into

three primary categories: thermal applications, electricity production, and chemical

processes. The most widely used applications are for water heating. Uptake of

electricity producing solar technologies is increasing for the application of

photovoltaic (primarily) and concentrating solar thermal-electric technologies. Due to

recent advances in solar detoxification technologies for cleaning water and air, these

applications hold promise to be competitive with conventional technologies. The

economics of solar energy has witnessed drastic improvements in the past few years.

This improvement came especially in the form of increased conversion efficiency, and

reductions in the cost of production.

1.9.1.2 Solar Energy Potential

There are different ways to reduce the cost of electricity in Kuwait. One of these ways

is to set up a Heat Rate improvement Program using the method that suits co-

28
generation plants. Also, Demand Side Management program should be set up to

recover some of the costs and to encourage conservation. Incremental Heat Rate must

be used to figure out which generating unit is most economical to serve the portion of

the load increase.

Also, the proposed electrical tie between the six Gulf States could play a big role in

saving these countries some future capital investments. A Power Pool should be

formed to determine the most economical way to serve any load in any area. Ways to

exploit the tremendous solar energy and other renewable energies like Fuel Cells

should be considered and studied for their feasibility in Kuwait. Power plants generate

adverse environmental and health impacts on the short, medium and long terms. Such

impacts need to be addressed, assessed and monitored compared to clean energy

resources. Climate change, rising energy prices and technology advances in renewable

energy utilization are all forces for reshaping the collective mindset of consumers,

turning many from "passive ratepayers" to highly informed, environmentally

conscious customers who want a role in using power. And now, with the emergence of

the technologies that make smart grids possible, companies can provide their

customers with the information and control they need to actually change their behavior

patterns and reduce usage and costs that show up on the utility bill. Many digital

companies are helping utilities by adding a layer of digital intelligence to their smart

grids.

These smart grids use sensors, meters, digital controls and analytic tools to automate,

monitor and control the two-way flow of energy. A power company can optimize grid

performance, prevent outages, restore outages faster and allow consumers to manage

energy usage right down to the individual networked appliance. Smart grids can also

29
incorporate new renewable energies such as solar and wind power, and interact locally

with distributed power sources, or plug-in electric vehicles.

The potential of solar energy resources is excellent in Gulf countries with an annual

global solar radiation varying between 4 to 8 kWh/m2. The region also enjoys high

direct normal radiation and low average cloud cover. The solar energy irradiated on

the

Fig. 1.6 Global horizontal and direct normal solar radiation [1.4]

ground is equal to 1-2 barrels of fuel oil per square meter and year. This magnitude of

resource can be used both in solar thermal applications and photovoltaic systems.

Thus, both distributed rural and centralized urban demand can be covered by solar

energy technologies. Fig. 1.6 shows the global horizontal solar radiation and direct

normal radiation in Kuwait compared to other countries. It is clear that Kuwait has a

high solar potential which can be utilized for electricity generation.

It is well known that Kuwait has no fresh water resources. However, Kuwait has an

abundance of solar energy capability. Solar energy use has been limited till now to

30
experimental, use and research purposes only. Solar energy could be used effectively

in different applications like Water Heating, lighting…etc. In other words, the use of

renewable energy is virtually very limited and needed to be thought of and considered

for supplying domestic loads.

Kuwait has been collaborating with local, regional, and international organizations and

companies to meet the new demands of power. Kuwait has gone above and beyond

establishing power supplies from various sources. Among these sources, Kuwait is

working on nuclear energy, and wind energy power. Having signed a 5 year deal with

Japan, for instance, nuclear power is estimated to make up 15- 20 percent of the total

electrical power supply. Another source Kuwait is considering to face power shortage

is solar and wind energy. As an example of innovative power resources is Solar

Power.

Because of high solar irradiation in the regions, State of Kuwait is resorting to buy

Solar PV Systems.

With the ministry of electricity and water estimate the demand for power to climb to

20,000MW by 2015 and 26,000MW in 2020, Kuwait plans to invest billions in

doubling its installed capacity through a number of projects by 2017, from currently

13000MW to 27,040MW.

As a means to relieve the power shortage in the foreseeable future, the government of

Kuwait has accepted a proposal to link up with the grids of other GCC countries. In

theory the grid linkup would provide each GCC country with additional spare capacity

to handle demand during summer. However, as observers have pointed out, peak

demand is the same throughout the GCC and without plans to provide greater spare

31
capacity the scheme seems unlikely to achieve its stated objective. Upon completion

of the project, Kuwait and Saudi Arabia will receive 1,200 MW of power capacity

each, UAE will receive 900 MW, Qatar 750 MW, Bahrain 600 MW and Oman 400

MW. The discrepancy between the amount of electricity generated and/or imported

and the amount consumed and/or exported is accounted for as loss in transmission and

distribution.

Owned by the government of Kuwait, the Sabiya facility is the largest power plant in

Kuwait and among the largest in the region. When the plant enters combined-cycle

operation by end 2012, the total plant output will be more than 2,000 megawatts. A

combined cycle plant dramatically boosts efficiency because waste heat is recovered

from the exhaust of the gas turbine, converted to steam and then fed to a steam turbine

to provide additional energy output.

The six gas turbines in the plant utilize emission-reducing technology that lowers
annual

Cos and NOx emissions by significant margins, which support Kuwait’s clean air

goals. In the hot arid region of the Arabian Peninsula including Kuwait, fresh water is

not available in adequate quantity in its natural form. Multi-stage flash (MSF) and

reverse osmosis (RO) are two desalination processes currently applied in different

countries to produce fresh water. These desalination processes need large amounts of

basic thermal and/or electrical energy for fresh water production from seawater or

underground brackish water. Primary energy requirement of MSF is thermal energy in

the temperature range of 80-120°C. A smaller amount of electrical energy is also

needed to operate miscellaneous pumps and controls. The RO system, unlike the MSF

system, needs only mechanical or electrical power. Solar energy can be effectively

32
used to produce the required low-grade thermal energy in a solar thermal or (PV)

system. The use of PV as the power source for pumping water is one of the most

promising areas in photovoltaic applications in Kuwait. With the increased use of

water pumping systems, more attention should be taken to their design and optimum

utilization in order to achieve the most reliable and economical operation. Thereby,

the water pumping system consists of the PV array, DC motor, centrifugal pump, a

storage tank that serves a similar purpose to battery storage and a maximum power

point tracker to improve the efficiency of the system. The pumped water is desired to

satisfy the domestic needs in remote areas in Kuwait.

1.9.2 Potential of Solar Power Generation in Kuwait

• Compatibility of Power Availability with National Power Demand o Power

demand in Kuwait is maximum during the summer months and during

sunshine hours. o Maximum power availability from solar systems is also

available during the same period.

• Plenty of Barren Land o Kuwait has plenty of desert land that is not being

used. o An area of 5 km2 or 0.03% of Kuwait land area is adequate to generate

500 MW solar power.

The increasing level of renewable energy penetration into public grid has recently

emphasized the importance of accurate renewable parameters (solar radiation, wind

speed..) forecasting to utilities and independent system operators (ISOs). System

operators are required to balance generation and load within a very tight band. This

means that if load is rising, or wind power is declining, other generation must rise to

33
keep the system balanced. Fig. 1.7 shows typical time frames that affect the operation

of the power system. Typically the most important time frames are for unit

commitment (days) and scheduling (daily). Large thermal units (e.g., those driven by

steam turbines) need long cold start times to synchronize them to the grid and start

generating electrical energy. Because of this, utilities and ISOs need to know a day-

ahead how much load is anticipated throughout the next day. With large penetrations

of renewable power, it also becomes essential to know with acceptable accuracy what

the total renewable power production on the system will be during each hour. When

renewable power forecasts are lower than what occurs, the system operator may have

to turn on additional units that

Fig. 1.7 Time frames of generation planning

were not needed, thereby increasing operating costs. When renewable power forecasts

are higher than what occurs, the system operator may not have committed enough

34
generation in advance and will either have to run expensive fast-start gas turbines or in

some cases shed load. The shorter term forecasts in the “load following” time frame

are also very important in balancing load, especially when renewable power may be

changing rapidly in magnitude either ramping up or down.

Accurate information on the meteorological data at Kuwait is very essential to the

development of renewable energy based projects. This information is used for project

design, in component layout, cost analysis and in calculations of the project overall

efficiency. In particular, the solar radiation, wind speed, the clearness, in addition to

other meteorological information such as humidity and, temperature, are extremely

important to assess the feasibility of utilizing renewable energy for electricity

generation.

1.10 Renewable Power Forecasting Techniques

Penn State/NCAR fifth-generation mesoscale model (MM5) 4D-Var system is the

most advanced system that was available to weather forecast [1.13], [1.15]. As shown

in

Figure 1.8, MM5 has been applied to assimilate zenith delay observations from

Global Positioning System satellites [1, 16] and to assimilate cloud-cleared brightness

temperatures from Geostationary Operational Environmental Satellite (GOES)

sounders [1, 17]. The MM5 produces very detailed and accurate simulations of

tropical cyclones when high-resolution and advanced physical parameterizations are

used [1, 13], [1, 17]. Due to current limitationof computer resources, weather control

4D-Var calculations were restricted to a horizontal grid spacing of 20 km.

35
The parameterizations of surface fluxes, radiated transfer, and cumulus convection are

currently limited in the MM5 4D-Var system. This technique can be applied to both

operational data and historical cases to aid in the scientific learning and pattern

recognition of forecasters. Renewable generation forecasts cover a wide range of time

frames. Typical power forecasts time frames include day-ahead hourly averages and

hourly or more frequent same day average forecasts. Power forecasts valid for periods

beyond several hours from the forecast origination time are developed by a two-step

process. First, Solar radiation and wind speed forecasts at renewable plant sites are

derived from numerical weather prediction model output. A wide range of models are

used by the various commercial firms as the basis for the speed predictions.

A recent trend in commercial forecasting is to use an ensemble of model output that

often includes the Weather Research and Forecasting model and large scale national

models such as the Global Forecast System. Ensemble results provide an indication of

the uncertainty of the renewable parameters and therefore power production forecasts

that are quite important to utility system operators. The forecasted solar radiation and

wind speeds must then be converted to forecasted power output in megawatts. This

information will then be used in utility operations such as for unit commitment and

dispatch of the conventional resources. This is not a straightforward process. The most

important problem is developing an accurate power production curve (the production

of energy by renewable resources.

36
 Fig. 1.8 Weather forcast using MM5 model [1.13]

In collaboration with the Army Test and Evaluation Command (ATEC), NCAR has

developed a Real-Time Four Dimensional Data Assimilation (RTFDDA) and

forecasting system. The RTFDDA system is built around the Penn State/NCAR

Mesoscale Model version 5 (MM5) and the Weather Research and Forecasting (WRF)

model. RTFDDA is capable of continuously collecting and ingesting diverse synoptic

and asynoptic weather observations from conventional and unconventional platforms,

and provides continuous 4-D weather analyses for accurate nowcasts and short-term

forecasts for mesoscale regions. Operational RTFDDA systems have been

implemented at seven ATEC test ranges and tens of other DoD, public and private

applications in the last seven years, providing rapidly updated, multi-scale weather

analyses and forecasts with the fine-mesh domain having 0.5 - 3 km grid increments.

The observational data ingested by the system includes WMO standard upper-air and

37
surface reports, wind profilers, satellite cloud-drift winds, commercial aircraft reports,

all available mesonet data, radar observations, and any special instruments that report

temperature, winds and moistures. Furthermore, the system has been expanded to

include the following new modeling and data assimilation capabilities that are highly

valuable for wind energy applications: a) Ensemble RTFDDA, which is a multi-

model, mesoscale data analysis and forecasting system that samples uncertainties in

the major components of RTFDDA and predicts the uncertainties in the weather

forecasts by performing an ensemble of RTFDDA analyses and forecasts; b) LES

(Large Eddy Simulation) modeling, which is nested down from the RTFDDA

mesoscale data assimilation and forecasts to LES models with grid sizes of ~100 m for

wind farm regions using GIS 30-m resolution terrain; and c) HRLDAS (High-

Resolution Land-Surface Data Assimilation System), which assimilates high-

resolution satellite vegetation and soil data to generate highresolution, accurate state of

soil moisture and temperature. This modeling system has been applied for supporting

wind energy prediction at wind farms operated by Xcel Energy. The modeling

capabilities will be described and demonstrated through case studies at selected wind

farms.

1.10.1 The Characteristics of Solar Radiation

The various solar energy applications/technologies are influenced by the character of

the resource, such as its directional nature (whether the sunlight is direct of diffuse —

by clouds, for example — and it’s angle of incidence on the collector surface), its

spectral nature (what specific wavelengths of sunlight the collector technology

38
responds to most effectively), and its variability. The variability characteristic can be

in the span of a few minutes (how clouds will affect power production), seasonal (how

climate patterns will affect the solar resource), inter-annual (how the resource will

vary year to year), or even decadal (how climate change could affect the resource).

This solar radiation computation program requires only a latitude value to predict

sunrise, sunset and sunshine period to compute the solar radiation equations. A new

predicted empirical equation has been added to this script to give a good agreement

and it has been tested for three different locations in the middle east which are Kuwait,

Riyadh and Jerusalem. The incident solar radiation values, based on monthly average

daily amounts can be obtained from the available references or the 22 years average

values which are recorded in NASA website [1.6].

Since the earth-sun distance varies each season, the apparent extraterrestrial solar

irradiation changes during the year. Therefore, the solar irradiation intensity depends

on the number of the day in the year [1.3]. The average daily extraterrestrial solar

irradiance is given by:

33048 360 n
S ET 1 0.034cos 365 cosL cossinHSR

180 HSR sinL sin

(0.1)

In clear sky atmosphere calculations, the clearness index Kt is defined as the ratio of the

horizontal insolation SH at the site to the ET insolation SET on a horizontal surface above

the site and just outside the atmosphere and is given by:

SH
Kt (0.2)
S ET

39
The relationship between the forecasted sky cover parameter SK and the global index,

GHI/GHIclear . Fig. 1.9 (a) is a plot of forecasted SK against measured global index for

all forecasts ranging from 4 to 8 hours [1.14]. Unlike the quasi-linear relationship

observed for the satellite-derived cloud cover CI (see Fig. 1.9 (b)). The relationship

between global irradiance index GHI index and cloud index CI represented by

groundobserved cloud cover is markedly non- linear. Kasten and Czeplak [1.5] had

proposed the following expression between the global index and ground-observed

cloud cover:

GHI GHI/ clear 1 0.75 N 3.4 (0.3)

8

where N is the ground-observed cloud cover in octas.

(a)

40
 (b)

Fig. 1.9 (a) Relationship between 4-8 hours SK forecasts and observed global
irradiance index (b) Relationship between satellite-derived cloud index CI and
observed global irradiance index

Another approach using clearance index is applied to represent the effect of cloud on

the global radiation. A high clearness index (Kt > 0.5) corresponds to clear skies,

where most of the SR will be direct beam, while a low one (Kt < 0.5) indicates

overcast conditions with diffuse SR. A monthly averaged clearness index for the site

of Kuwait can be obtained from National Aeronautics and Space Administration

(NASA) website for Kuwait area over 22 years (1983 – 2005) [1.6], as given in Table

1.7.

Table 1.7: Monthly averaged clearness index in Kuwait area

Latitude 29.33o
22 Year Average Kt
Longitude 47.5o
January 0.52
February 0.57
March 0.56
April 0.56
May 0.62

41
 June 0.69
July 0.68
August 0.67
September 0.65
October 0.60
November 0.50
December 0.47
Annual average 0.59

From the best curve fitting for the values of Kt in Table 1, the clearness index can be

expressed to a good accuracy as a function of day number n, as

Kt 3.807 10 62
n 0.001124n 0.6139 (0.4)

The daily global SR on horizontal surfaces, SH, in Wh/m2, can be found from (1.2)

The hour by hour SR data is essential if the radiation on the horizontal surfaces is to be

estimated. We can follow the same procedure as for the daily radiation except for

expressing the hour angle in terms of solar time in hours (local apparent time, LAT) as

in:

H 15 (t 12) (0.5)
and letting the solar time t vary between sunrise time and sunset time according to the

longitude of site. For the Kuwait area with a longitude of 47.5º, the standard sunrise

and sunset times are 6:00 am and 18:00 pm, respectively, and ET S may be obtained

from Kuwait Metrological Centre.

The consumption of electrical energy in Kuwait indicates that almost one half of the

generated energy in Kuwait is consumed for domestic purposes. At the same time,

most of the load in residential, industrial, and commercial areas is due to air

conditioning (A/C) systems.

42
According to the maximum and minimum load distributions shown in Table 1.3 for

the year 2010, Kuwait electrical load is characterized by high load in summer and low

load in winter depending on the increase and decrease in the values of temperatures

and relative humidity.

The meteorological data for Kuwait, latitude angle (L = 29.33º), longitude angle ( γ =

47.5o ), clearness index Kt (almost cloudless atmosphere for nine months), solar time t

(varying from 6:00 to 18:00), and the corresponding hour angle H, has been used in

the model and the intensity of daily and hourly global SR on horizontal surfaces SH in

Kuwait area has been obtained as a function of both day number n (with a day number

increment of one day) and time (with a time increment of one hour). These results can

be obtained for any day of the year and for any hour of the day.

1.11 Solar cooling

Today, there are two main solar air-conditioning technology options: solar thermal

airconditioning, where the solar absorption cooling is the first type of this option and it

is still practical for remote building in places where there is an excess of heat energy

available. Another option is the solar photovoltaic air-conditioning, by using

electricity from renewable sources to power the conventional cooling equipment. ‘‘On

the other hand, the market introduction of photovoltaic systems is much more

aggressive than that of solar thermal power plants; cost reductions can be expected to

be faster for photovoltaic systems. [1.7].

The cooling demand generally follows the outside ambient air temperature. Normally,

higher temperature leads to higher cooling load due to a heat transition through the

building’s envelope from hotter outside to a cooler inside of the building.

43
Solar air-conditioning systems can be divided into two groups of systems:‘‘ solar

autonomous systems and solar-assisted systems. In a solar autonomous system “all”

energy used by the air-conditioning system is solar energy. In a solar assisted system

the solar energy covers a certain fraction of the energy used by the air-conditioning

system and the rest of the energy is provided through an auxiliary or backup system’’

[1.7]. Only the solar-assisted systems are considered in this study.

Solar air-conditioning technologies are any air-conditioning system that use solar

radiation as source of power to drive the cooling process in order to produce cold air

for buildings. This can be achieved through solar thermal conversion or solar

photovoltaic (PV) conversion. In PV air-conditioning systems, PV cells arranged in

modules convert solar radiation to electric power which then drives a traditional

compression chiller.

Solar thermal air-conditioning technology converts solar radiation to heat power

(through thermal collector) which is fed into a thermal cooling process or into a direct

air-conditioning system.

1.11.1 Solar PV Air -Conditioning

PV air-conditioning system consists of three main parts: solar energy collection

(includes PV cells) which converts solar radiation into electric power in order to drive

the electric machine heat pump. This machine is any electric traditional air-

conditioning system which converts the electric power to cooling power. The cooling

power is distributed for space cooling either directly in a decentralized system or by a

cooling coil and sometimes by a hidronic system [1.8]. The PV air-conditioning

system can include a storage system (battery system) or it can be without a battery

44
system. Thermal air-conditioning systems can be divided in two groups [1.8]: a solar

heat driven air-conditioning system which consists of a solar thermal collectors (high

temperature or medium temperature hot water) where typically flat plate collectors,

evacuated tube collectors or concentrated parabolic collectors are used. This converts

the solar radiation to heat power. Then this power is provided to drive another system

in order to produce cooling power for buildings.

This system can be either an electrical air-conditioning (such as traditional

airconditioning) or thermal air-conditioning (such as absorption/adsorption chillers) or

both. The other type is a desiccant cooling system which uses water as refrigerant in

direct contact with air and the desiccant dehumidification is combined with an

additional cooling system which may be a conventional cooling coil or evaporative

cooling.

The PV air-conditioning without a storage scenario consists of a PV–module, inverter,

a compressed chiller and a system distribution. As shown in Fig. 1.10, the system set

up of a PV air-conditioning with storage is similar to the first scenario, but it

additionally has a storage (battery) system and a charge controller.

The PV module converts solar radiation into electric power as direct current (DC ).

The solar charge controller regulates the voltage and the current which comes from the

PV module into the battery. This prevents from the overcharging of the battery and

increases the battery life.

The inverter converts DC into alternating current (AC ) which is needed to drive the

compressed chiller. The battery stores the excess energy for supplying the compressed

chiller when there is no enough solar radiation to cover the cooling demand. The

compressed chiller converts the AC power to the cold air. The compressed chiller is

45
supplied as a back-up with an electric AC power from the grid, when there is not

enough DC power from the PV-array and the battery bank, especially at night, evening

and morning of the day when there is no enough solar radiation to drive the

compressed chiller. The first step in designing a solar PV air-conditioning system is to

find out the cooling power and energy consumption that need to be supplied by the

PV-array.

Fig. 1.10 Schematic flow diagram for solar PV air-conditioning with storage.
Energy produced by the PV array is accumulated and stored in batteries for use on

demand. The battery system is designed for a PV air-conditioning with storage. In this

scenario, when the electric DC power is higher than the electric power needed by the

compression chiller. The battery accumulated and stored the excess DC power as

electric energy. The solar charge controller regulates the voltage and the current which

comes from the PV module into battery [1.9].

46
When there is no enough electric power coming from the PV-array and the battery

system to cover the cooling power demands, the back-up system is designed to deliver

the remaining needed AC power. The back-up system is assumed to have a direct

connection to the electricity grid and to the compressed chillers. This connection is

considered and designed for both PV air-conditioning scenarios without and with

storage [1.10].

1.11.2 Solar Thermal Air-conditioning

The thermally driven air-conditioning process is the heart of every solar cooling

system. Thermally driven air-conditioning systems are available on the market which

commonly utilize sorption processes. For air-conditioning applications, mainly

absorption chillers using the sorption pair Lithium bromide-water (LiBr-H2O) are

applied [1.11] since they require a comparatively low temperature as heat input. Most

of the thermally driven cooling system and solar assisted air-conditioning systems

installed today are based on absorption chillers [1.11].

In this solar thermal system, an absorption chiller coupled with a solar heating system

and an auxiliary energy supply as back-up electric heater is analyzed. In the assumed

case, the collector converts the solar radiation into heat and then the pump delivers it

to the storage tank. The storage tank then supplies the absorption chiller with thermal

energy to produce cold water. The coil and the fan system transfer the cooling power

from cold water to the inside air of the building. A duct system distributes it in the

building. If solar heat is insufficient e.g. at night or during cloudy days, the

conventional back-up electric heater is connected directly with the storage tank can

47
provide heat. The system components and design can be described as follows in Fig.

1.11:

Fig. 1.11 Solar thermal air-conditioning system scenario, coupling of an absorption

chiller with a solar heating system.

A solar collector is a special kind of heat exchanger converting the solar radiation into

thermal heat which is carried by a working fluid, e.g. water which is flowing through

the collector. There are three types of solar collectors which are typically used in a

solar thermal air-conditioning systems: flat plate collectors, evacuated tube collectors

and concentrated parabolic collector.

In the solar thermal air-conditioning system, the temperature level that should be

supplied by the solar thermal collector depends on the cooling technology used. Flat

plate collectors can be designed for applications requiring energy delivery at moderate

temperatures, up to perhaps 100 0C above the ambient temperature. The operation of

absorption air-conditioning with energy from flat plate collectors and storage systems

is the most common approach to solar cooling.

48
Flat plate collectors are fixed and there is no tracking system. The collectors should be

oriented directly towards the equator. The collector’s location in the northern

hemisphere should be facing the south and vice-versa in order to maximize the amount

of daily and seasonal solar energy received by the collector. The optimal tilt angle of

the collector is an angle equal to latitude of its location [1.11]. However, in summer

the tilt angle should be smaller than the latitude to receive more solar radiation.

The solar photovoltaic air-conditioning system stores the excess DC in batteries.

Similarly, It is necessary to use a thermal energy storage tank, either heat or cold

storage in thermal air-conditioning system.

These two alternatives are not equivalent in capacity, costs or effect on the overall

system design and performance. The required capacity of a cold storage tank is less

than that required of a heat storage tank because the heat storage has higher

conversion efficiency than the cold storage tank. In addition, heat storage tanks can be

used for other applications for example domestic hot water or space heating.

When there is no enough solar radiation (e.g. at evening, night or on cloudy days), it

will be a necessary to have a back-up system for the cold production or to allow the

solar air-conditioning system to continue.

Two different back-up approaches can be used to achieve this objective, either back-up

heating or cooling systems. The back-up heating system usually uses burners (oil, gas

or pellet) or electric heater connected directly to the heat storage tank. The back-up

cooling system usually uses conventional vapor pressure cooling devices.

One of the main objectives of the solar air-conditioning systems is to save primary

energy consumption. Recent studies have indicated that PV air-conditioning scenario

49
with storage behaves and compensates the cooling demand better than the solar

thermal air-conditioning with storage scenario and needs less storage to cover the

same amount of cooling load demand. However, the storage system in the PV air-

conditioning scenario is minor and the direct compensation is major. Therefore, the

PV performance has to be analyzed under different weather condition in Kuwait.

1.11.3 District cooling

The extensive use of air conditioning for indoor cooling in offices and large

commercial buildings in Kuwait represents a major part of the power and electricity

consumption in that country. In Kuwait approximately 45% of the total annual

exported electrical energy is consumed solely by air conditioning systems as a result

of the very high ambient temperatures occurring between March and October.

Furthermore, it was estimated that air conditioning systems represent about 62% of the

peak electrical load [1.12].

This was motivated by the extreme climatic conditions in Kuwait and the necessity to

reduce both maximum power demand and energy consumption whilst being

economically feasible. To assess the potential of cool thermal storage, the air

conditioning system for a typical medium size building in Kuwait, was studied with

and without a cool thermal storage system. The recent studies demonstrate that

internal iceon-coil and chilled water storage systems are suitable storage technologies

that can be implemented in Kuwait.

Air and water-cooled water chillers are commonly used for AC of medium and large

sized buildings in Kuwait. Air-cooled chillers are generally favoured for medium size

systems and water-cooled for large systems. In these systems, the chillers cool the

50
water, and the cooled water is pumped through the piping system to the various air

handling and fan coil units in the building. The fan in the air handling and coil units

circulates the air in the ducting system and within the conditioned space. The return air

from the room is then cooled and dehumidified by the cooling coils and returns again

to the room.

Basically, in this process, the circulated chilled water in the cooling coil picks up the

heats that is gained by the room, supply and return ducts, and fan motors, and is

returned to the chiller for cooling in the evaporator. The refrigerant in the evaporator

absorbs the heat of the return chilled in the chiller as it evaporates at a low temperature

and pressure, and gives up this heat to the atmosphere as it condensates at a higher

temperature and pressure.

In an air-cooled system, the power demand of the chiller is a strong function of

building load and ambient dry bulb temperature. However, the power demand of the

AHU and chilled water pumps (if constant speed motors are used in the system) are

generally constant during the whole summer period irrespective of the building

cooling load and weather conditions. In water-cooled systems the main components in

the AC system that consume electricity are similar to those of the air-cooled systems

but with the addition of the condenser water pumps and motors of the cooling tower

fans. MEW promotes the implementation of District Cooling in the new five cities

coming up in Kuwait. One of them is a commercial city and the rest are residential

cities. District

Cooling is the concept of centralized production and distribution of cooling energy,

using chilled water delivered via an underground insulated pipeline to office,

industrial and residential buildings to cool indoor air within a district. The district

51
cooling is fast becoming a reality in the GCC states, which are suffering from high

growing because of the need for comfort cooling, non-ozone depleting coolants, and

cutting energy consumption.

By district cooling, the solar energy is gained through the collector and accumulated in

the storage tank. Then, the hot water in the storage tank is supplied to the generator to

boil off water vapour from a solution of Lithium Bromide and water. The water

vapour is cooled down in the condenser and then passed to the evaporator where it

again is evaporated at low pressure, thereby providing cooling to the required space.

Meanwhile, the strong solution leaving the generator to the absorber passes through a

heat exchanger in order to preheat the weak solution entering the generator. In the

absorber, the strong solution absorbs the water vapor leaving the evaporator.

Cooling water from the cooling tower removes the heat by condensation. Since the

temperature of the absorber has a higher influence on the efficiency of the system than

the condensing temperature, the heat rejection (cooling water) fluid, is allowed to flow

through the absorber first and then to the condenser. An auxiliary energy source is

provided, so that hot water is supplied to the generator when solar energy is not

sufficient to heat the water to the required temperature level needed by the generator.

More than 70 percent of the domestic energy consumption is spent on comfort cooling.

By 2030, the demand will grow to 32,000 MW. Kuwait uses 350,000 barrels of oil per

day to generate electricity, and in the next 15 years, this figure will climb to 1 million

bpd. District cooling can cut power consumption by up to 30 percent.

The contribution of this report is to find solutions to the rationing of energy use with

the aim to preserve Kuwait’s oil wealth and to priorities future needs during the

52
gradual shift towards the use of alternative and renewable energy in power generation,

with emphasis on solar energy and PV generation. Solar energy as a renewable energy

resource will play a major role in Kuwait’s future energy supply. Kuwait has an

abundance of solar energy capability [1.12].

1.12 Meteorological Data for Kuwait

To study the utilization potential of solar energy in Kuwait the meteorological data

will be further explained in next section. These data sets were used to assess the

performance of the PV panel and generate the results presented in this study.

The data file was hourly data included the horizontal solar radiation (beam, diffuse and

global), ambient air temperature and relative humidity [1.6]. Figs. 1.12 to 1.20

represent the horizontal global irradiation, ambient air temperature, cloud cover,

participation and relative humidity data for Kuwait.

1.12.1 Temperature

The hottest day of 2011 was July 9, with a high temperature of 50°C. The hottest

month of 2011 was July with an average daily high temperature of 46°C.

53
 Fig. 1.12 Temprature variation in Kuwait

1.12.2 Cloud Coverage

The fraction of time spent in each of the five sky cover categories over the course of

2011 on a daily basis. From top (most blue) to bottom (most gray), the categories are

clear, mostly clear, partly cloudy, mostly cloudy, and overcast. Pink indicates missing

data. The bar at the top of the graph is gray if the sky was cloudy or mostly cloudy for

more than half the day, blue if it is clear or mostly clear for more than half the day,

and blue-gray otherwise.

The cloudiest month of 2011 was June, with 27% of days being more cloudy than
clear.

The longest spell of cloudy weather was from February 20 to February 25, constituting

6 consecutive days that were cloudier than they were clear.

54
 Fig. 1.13 Cloud Coverage in year 2011

1.12.2.1 Hourly Cloud Coverage

The full year of hourly cloud coverage reports with the days of the year on the

horizontal and the hours of the day on the vertical. The sky cover is color coded: from

most blue to most gray, the categories are clear, mostly clear, partly cloudy, mostly

cloudy, and overcast. Pink indicates missing data.

Fig. 1.14 Hourly Cloud Coverage

55
1.12.3 Precipitation

This station reports the quantity of liquid precipitation but not observations of weather

events in the area. For this reason, this section will only discuss the quantity of

precipitation and not its type.

Fig. 1.15 Precipitation amount during 2011

1.12.4 Humidity

Humidity is an important factor in determining how weather conditions feel to a

person experiencing them. Hot and humid days feel even hotter than hot and dry days

because the high level of water content in humid air discourages the evaporation of

sweat from a person's skin.

When reading Fig. 1.16, keep in mind that the hottest part of the day tends to be the

least humid, so the daily low (brown) traces are more relevant for understanding

daytime comfort than the daily high (blue) traces, which typically occur during the

night. Applying that observation, the least humid month of 2011 was June with an

56
average daily low humidity of 6%, and the most humid month was January with an

average daily low humidity of 37%.

But it is important to keep in mind that humidity does not tell the whole picture and

the dew point is often a better measure of how comfortable a person will find a given

set of weather conditions. Please see the next section for continued discussion of this

point.

Fig. 1.16 Humidity Variation for year 2011

The daily low (brown) and high (blue) relative humidity during 2011 with the area

between them shaded gray and superimposed over the corresponding averages (thick

lines), and with percentile bands (inner band from 25th to 75th percentile, outer band

from 10th to 90th percentile).

57
1.12.5 Dew Point

Dew point is the temperature below which water vapor will condense into liquid
water.

It is therefore also related to the rate of evaporation of liquid water. Since the evaporation

of sweat is an important cooling mechanism for the human body, the dew point is an

important measurement for understanding how dry, comfortable, or humid a given set of

weather conditions will feel.

Generally speaking, dew points below 10°C will feel a bit dry to some people, but

comfortable to people accustomed to dry conditions; dew points from 10°C to 20°C

are fairly comfortable to most people, and dew points above 20°C are increasingly

uncomfortable, becoming oppressive around 25°C.

To take some examples from Fig. 1.17, and basing our categorization on the daily high

dew point in 2011, January had 19 dry days, 12 comfortable days, and no humid days;

April had 11 dry days, 18 comfortable days, and no humid days; July had 15 dry days,

10 comfortable days, and 6 humid days; and October had 8 dry days, 14 comfortable

days, and 9 humid days.

58
 Fig. 1.17 Dew Point Variation during 2011

The daily low (blue) and high (red) dew point during 2011 with the area between them

shaded gray and superimposed over the corresponding averages (thick lines), and with

percentile bands (inner band from 25th to 75th percentile, outer band from 10th to

90th percentile).

1.12.6 Wind

The highest sustained wind speed was 16 m/s, occurring on April 4; the highest daily

mean wind speed was 7 m/s (June 16); and the highest wind gust speed was 21 m/s

(April 4). As indicated in Fig. 1.18, the windiest month was March, with an average

wind speed of 4 m/s. The least windy month was December, with an average wind

speed of 2 m/s.

59
 Fig. 1.18 Wind Speed Variation

The daily low and high wind speed (light gray area) and the maximum daily wind gust

speed (tiny blue dashes).

60
 Fig. 1.19 Sunshine Duration

Fig. 1.20 Diffuse Radiation

1.12.7 Measurement of meteorological data

The site of the measurement station was located at Kuwait International Airport

(29.217_Latitude N, 47.983_Longitude E). The meteorological data were collected

stored to calculate the mean, maximum and minimum hourly values of these data.

From the hourly data set, daily and monthly statistics were made for the stored

meteorological data. As an example, Figs. 1.19 and 1.20 display the monthly sunshine

duration and solar radiation.

61
1.12.8 Global Solar Radiation

From the recorded meteorological data it is clear that the daily average and maximum

global radiation as well as temperatures are higher in summer time and lower in the

winter. Fig. 1.21 describes the daily average global solar radiation for the whole

period of 2007-2011. The graph shows that the daily highest global radiation of 2650

W/m2 was recorded in July, while the lowest daily average solar radiation of 930 W/

m2 was recorded in February. Fig. 1.21 also shows downward excursions in winter,

especially in January and February. These excursions might be due to sandstorms, rain

events and higher air mass in winter. The higher air mass in winter somewhat reduces

the clear sky data by the absorption along the longer path length. The measurements

agreed with the

22-year average global solar radiation data of the NASA SSE model [1.6].

Fig. 1.21 Daily Variation of Global solar radiation from 2007-2011

62
1.12.9 Wind Energy conversion

The feasibility of wind energy conversion is under investigation by the Kuwaiti

Institute for Scientific Research (KISR). The main objective of this study is to develop

sustainable building with integrated renewable energy systems, in particular wind

power. Fig. 1.22 shows a sketch of how the power output from a wind turbine varies

with steady wind speed.

Fig. 1.22 Wind Turbin Characterstics

1. Cut-in speed

At very low wind speeds, there is insufficient torque exerted by the wind on the

turbine blades to make them rotate. However, as the speed increases, the wind turbine

will begin to rotate and generate electrical power. The speed at which the turbine first

starts to rotate and generate power is called the cut-in speed and is typically between 3

and 4 meters per second.

2. Rated output power and rate output wind speed

63
As the wind speed rises above the cut-in speed, the level of electrical output power

rises rapidly as shown. However, typically somewhere between 12 and 17 metres per

second, the power output reaches the capacity limit of the electrical generator. This

limit to the generator output is called the rated power output and the wind speed at

which it is reached is called the rated output wind speed. At higher wind speeds, the

design of the turbine is arranged to limit the power to this maximum level and there is

no further rise in the output power. How this is done varies from design to design but

typically with large turbines, it is done by adjusting the blade angles so as to keep the

power at the constant level.

3. Cut-out speed

As the speed increases above the rate output wind speed, the forces on the turbine

structure continue to rise and, at some point, there is a risk of damage to the rotor. As a

result, a braking system is employed to bring the rotor to a standstill. This is called the

cut-out speed and is usually around 25 metres per second.

By comparing the recorded wind speed with the wind power curve results in a

significant number of days per year when wind speed is less than 4 m/s and hence the

wind power would drop to zero. Therefore, the potential of utilizing wind energy is

generally lower than that of solar energy. This creates a serious problem in planning a

nation's electricity supply system. In Kuwait, the government target is to increase the

share have of electricity supply by provided by renewable sources - which will

effectively all be provided by solar energy. In the following sections characteristics of

Photovoltaic cells are studied under the climate conditions of Kuwait.

64
References

[1.1] Statistical Year Book 2010 - Electrical Energy - Ministry of Electricity &
Water
[1.2] Ali Hajiah," Sustainable Enrgy in Kuwait: Challenges and opportunities",
UNDP Regional Meeting, 6 Oct. 2010.
[1.3] F. Q. Al-Enezi and J. K. Sykulski, Modeling of a Photovoltaic Module
Considering the Solar Energy Available from Horizontal Surfaces over Kuwait
Area, JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY, VOL.
10, NO. 2, JUNE 2012
[1.4] Renewable Energy Applications and Potentials in MENA Region, National
Links with EU or Other Organizations: UNDP
[1.5] Kasten, F., and Czeplak, G., 1980, “Solar and Terrestrial Radiation Dependent
on the Amount and Type of Cloud,” Sol. Energy 24_2_, pp. 177–189.
[1.6] NASA SSE Model, 1983−2005, Atmospheric Science Data Center.
[1.7] Dr. Volker Quaschning, Dr. Manuel Blanco Murie, Photovoltaics or Solar
Thermal Power Plants?, DLR, Plataforma Solar de Almería, Spain, CIEMAT,
Plataforma Solar de Almería, Spain, VGB Congress Power Plants 2001 •
Brussels • October 10 to 12, 2001, the contents available at :
http://volkerquaschning.de/downloads/VGB2001.pdf.
[1.8] M.Eltawil and D.Samuel “Vapour Compression Cooling System Powered By
Solar PV Array for Potato Storage" Agricultural Engineering International:
the CIGR Ejournal. Manuscript EE 06-003. Vol. IX. June, 2007
[1.9] Al-Salaymeh, Al-Hamamre, Sharaf, Abdelkader, a Technical and economical
assessment of the utilization of photovoltaic systems in residential buildings:
The case of Jordan, December 2009, contents lists available at Science Direct
Website, www.sciencedirect.com.

[1.10] Hining ,H.-M(Ed): Solar-Assisted Air – Conditioning in Building –A

Handbook for planners. Springer-Verlag/wien, 2004.
[1.11] Nada Mekki, Application of absorption solar air-conditioning Technology
depending on climate and building standard, Energy and semiconductor
Research Laboratory, Department of physics, Faculty of Mathematics
&Science ,Carl Von Ossietzky University, Oldenburg/F.R. Germany, 2008.
[1.12] Mohamed Sebzali, Assessment of Cool Thermal Storage Strategies in Kuwait,
Ph.D Theses, Cranfield University, 2007.
[1.13] Grell et al. 1995, A Description of the Fifth-Generation Penn State/NCAR. [1.14]
R. Perez, S. Wilcox, D. Renne, K. Moore, A. Zelenk, “ Forecasting Solar Radiation:
Preliminary Evaluation of an Approach based upon the National Forecast Data Base”,
Solar Energy, Vol. 81, Issue 6, June 2007, pp. 809-812.

65
[1.15] M. De Pondeca, X. Zou, “ A Case Study of Variation Assimilation of GPS
Zenith Delay Observation into a Mesoscale Model”, J. Appl. Meteoro., Vol.
40, 2001, pp. 1559-1576.
[1.16] X. Zou, Q. Xiao, A. Lipton, Q. Modica, “ A Numerical Study of the Effect of
GOES Sounder Cloud-Cleared Brightness Temperature on the Prediction of
Hurricane Felix”, J. Appl. Meteoro., Vol. 40, 2001, pp. 34-35.
[1.17] Y. Lui, D.-L. Zhang, M. Yau, “ A Multi-Scale Numerical Study of Hurricane
Andrew 1992- Part II: Kinematics of Inner-Core Structure”, Mon. Weather
Rev., Vol. 127, 1999, pp 2597-2616.

66
 Chapter 2 Design and Modeling of Photovoltaic Systems

2.1 Definition

A photovoltaic system consists generally of one or more solar panels to convert solar

energy into electricity. It consists of multiple components, including the photovoltaic

modules, mechanical and electrical connections and mountings as well as electronic

devices to regulate and/or modify the electrical output [2.1].

2.2 Photovoltaic Arrangements

2.2.1 Photovoltaic Cell

PV cells are made of semiconductor materials, such as silicon. For solar cells, a thin

semiconductor wafer is specially treated to form an electric field, positive on one side

and negative on the other. When light energy strikes the solar cell, electrons are

knocked loose from the atoms in the semiconductor material. If electrical conductors

are attached to the positive and negative sides, forming an electrical circuit, the

electrons can be captured in the form of an electric current and generate electric

power. This electric power can then be used to power a load [2.1]. A PV cell can

either be circular or square in construction.

2.2.2 Photovoltaic Module

Due to the low voltage generated in a PV cell (around 0.5V), several PV cells are

connected in series (for high voltage) and in parallel (for high current) to form a PV

module for the desired output. Separate diodes may be needed to avoid reverse

currents, in case of partial or total shading, and at night. The p-n junctions of mono-

67
crystalline silicon cells may have adequate reverse current characteristics and these are

not necessary. Reverse currents can cause power loss and also lead to overheating of

shaded cells. Solar cells become less efficient at higher temperatures and in extreme

cases ventilation is required for solar panels [2.2].

2.2.3 Photovoltaic Array

The power that one module can produce is not sufficient to meet the requirements of

home or business. Most PV arrays use an inverter to convert the DC power into

alternating current that can supply loads such as motors, lights etc. The modules in a

PV array are usually first connected in series to obtain the desired voltages; the

individual modules are then connected in parallel to allow the system to produce more

current

[2.2]. The different PV system configurations are displayed in Fig. 2.1.

Fig. 2.1 Photovoltaic systems

68
2.3 Materials Used in PV Cells

The materials used in PV cells are classified as follows:

2.3.1 Single-Crystal Silicon

Single-crystal silicon cells are the most common in the PV industry. The main

technique for producing single-crystal silicon is the Czochralski (CZ) method. High-

purity polycrystalline is melted in a quartz crucible. A single-crystal silicon seed is

dipped into this molten mass of polycrystalline. As the seed is pulled slowly from the

melt, a singlecrystal ingot is formed. The ingots are then sawed into thin wafers about

200-400 micrometers thick. The thin wafers are then polished, doped, coated,

interconnected and assembled into modules and arrays [2.2].

2.3.2 Polycrystalline Silicon

Consisting of small grains of single-crystal silicon, polycrystalline PV cells are less

energy efficient than single-crystalline silicon PV cells. The grain boundaries in

polycrystalline silicon hinder the flow of electrons and reduce the power output of the

cell. A common approach to produce polycrystalline silicon PV cells is to slice thin

wafers from blocks of cast polycrystalline silicon. Another more advanced approach is

the “ribbon growth” method in which silicon is grown directly as thin ribbons or

sheets with the approach thickness for making PV cells [2.2].

2.3.3 Gallium Arsenide (GaAs)

A compound semiconductor made of two elements: Gallium (Ga) and Arsenic (As).

69
GaAs has a crystal structure similar to that of silicon. An advantage of GaAs is that it

has high level of light absorptivity. To absorb the same amount of sunlight, GaAs

requires only a layer of few micrometers thick while crystalline silicon requires a

wafer of about 200-300 micrometers thick. Also, GaAs has much higher energy

conversion efficiency than crystal silicon, reaching about 25 to 30%.The only

drawback of GaAs PV cells is the high cost of single crystal substrate that GaAs is

grown on [2.2].

2.3.4 Cadmium Telluride (CdTe)

It is a polycrystalline compound made of cadmium and telluride with a high light

absorbility capacity (i.e a small thin layer of the compound can absorb 90% of solar

irradiation).The main disadvantage of this compound is that the instability of PV cell

or module performance. As it a toxic substance, the manufacturing process should be

done by extra precaution [2.1, 2.2].

2.3.5 Copper Indium Diselenide (CuInSe2)

It is a polycrystalline compound semiconductor made of copper, indium and selenium.

It delivers high energy conversion efficiency without suffering from outdoor

degradation problem. It is one of the most light-absorbent semiconductors. As it is a

complex material and toxic in nature so the manufacturing process face some problem

[2.2].

70
2.4 Characteristics of PV Cell

A photovoltaic cell is simply represented by a forward biased diode of large area with

a photo-voltage. The photo-voltage is created from the dissociation of electron hole

pairs created by incident photons within built-in field of the junction [2.3, 2.4]. In the

dark, the solar cell is not an active device; it works as a diode (p-n junction). However,

if it is connected to an external supply (large voltage) it generates a current ID, called

the diode current or dark current. Fig. 2.2 illustrates the I-V characteristic of the solar

cell for a fixed ambient irradiation G and a given cell temperature Tc.

If the cell's terminals are connected to a variable resistance R, the operating point is

determined by the intersection of the I-V curve of the solar cell with the load I-V

curve. For a resistive load, the load characteristic is a straight line with a slope

I/V=1/R. Note that the power delivered to the load depends on the value of the

resistance only. However, if the load R is small, the cell operates in region MN of the

curve, where the cell behaves as a constant current source, almost equal to the short

circuit current. On the other hand, if the load R is large, the cell operates on the region

PS of the curve, where the cell behaves more as a constant voltage source, almost

equal to the open circuit voltage.

71
 Fig. 2.2 Characteristic of a PV cell

A solar cell can be characterized by the following fundamental parameters as indicated

in Fig. 2.2.

Short Circuit Current Isc

It is the greatest value of the current generated by a cell. It is produced under short

circuit conditions (V=0). The short circuit current can be considered equivalent to the

photocurrent (Isc = Iph).

Open Circuit Voltage Voc

This corresponds to the voltage drop across the diode (p-n junction), when it is

traversed by the photocurrent Iph (ID = Iph), when the generated current is I = 0. It

reflects the voltage of the cell at night and can be expressed as:

Voc BKTc ln I ph (0.6) q Io

72
where, B is the ideality factor, K is Boltzmann's constant, Tc is the absolute

temperature of the cell, q is the electronic charge and Iph is the photocurrent. Io is the

dark saturation current and it is strongly dependent on temperature [2.1, 2.2].

Maximum Power Point Pmax

It is the operating point A (Vmax, Imax) in Fig. 2.2, at which the power dissipated in the

resistive load is maximum:

Pmax Imax V max (0.7)

Maximum Efficiency η

It is the ratio between the maximum power and the incident light power:
Pmax I max V max (0.8)
Pin Ac G

where, G is the ambient radiation and Ac is the cell area.

Fill Factor FF

It is the ratio of the maximum power that can be delivered to the load and the product

of Isc and Voc:

FF Pmax Imax Vmax (0.9)

Voc Isc Voc Isc

The fill factor is a measure of the real I-V characteristic. Its value is higher than 0.7 for

good cells. The fill factor diminishes as the cell temperature is increased.

73
2.4.1 PV Module

Cells are normally grouped into modules in order to meet specified power output

requirements as previously shown in Fig. 2.1. The cells are encapsulated with various

materials to protect them and the electrical connectors from the environment. Fig. 2.3

represents how I-V characteristics are modified in the case when two identical cells

are connected in series or in parallel [2.2].

8 1
2
Two
Cells

I (A) Two I (A)

One Cells
One
Cell Cell

0 0
0 45 0
V V
(V) (V)

(a) (b)
Fig. 2.3 I-V characteristics of two identical PV cells. (a) In series. (b) In parallel.

Fig. 2.4 PV array

2.4.2 PV Array

The modules in a PV system are typically connected in arrays. Fig. 2.4 illustrates the
array configuration with Mp parallel branches each with Ms modules in series. If it is

74
assumed that the modules are identical and the ambient irradiation is the same on all
the modules then, the array's terminal voltage can be determined by:
VA M VS M (0.10)

and the array's current is:

IA M P IM (0.11)
2.4.3 Effect of irradiance and temperature on the module

characteristics

The photocurrent Iph is proportional to solar irradiance. The open circuit voltage

increases slightly with increasing irradiance, while the short circuit current and the

optimal power increase linearly with irradiation increase. Temperature has a

significant effect on the solar cell I-V characteristics. The photocurrent will increase

slightly with increasing cell temperature. The open circuit voltage decreases linearly

with increasing cell temperature owing to the exponential increase in saturation

current [2.3, 2.4]. These effects will be discussed in details later in the report.

2.5 Types of PV Systems

There are two basically different PV systems: those with a connection to an (available)

electric grid and remote or "stand-alone" systems. While in the first case the grid

serves as an ideal storage component and ensures system reliability, the stand-alone

systems require a storage battery. This battery serves as a buffer between the

fluctuating power generated by the PV cells and the load. In order to ensure

continuous power supply, even under extreme conditions, a back-up generator is often

also installed. Buildingintegrated PV systems have an economical advantage over

conventional PV generator systems: The PV modules serve for multiple purposes.

They are part of the building envelope, ideally replacing conventional facade or roof

75
material. Modem commercial building facades often cost as much as a PV façade

which means immediate or shortterm payback for the PV system. Depending on the

type of integration, the PV modules may also provide shading or noise protection.

Here again, the costs for replaced conventional means for these purposes may be

deducted from the initial PV costs.

2.5.1 Stand-Alone Systems

PV systems are most effective at remote sites off the electrical grid. Their high

reliability and low servicing requirements make them ideally suited for applications at

(for parts of the year) unattended sites. The costs for a PV system compete in this case

against the cost for a grid connection or other possible ways of remote energy supply.

As stated above, a storage battery is needed. Excess energy produced during times

with no or low loads charges the battery, while at times with no or too low solar

radiation the

76
 Fig. 2.5 Block diagram of a stand-alone PV system

loads are met by discharging it. A charge controller supervises the charge/discharge

process in order to ensure a long battery lifetime. As in the grid-connected systems, an

inverter, when required, transforms DC to AC electricity as given in Fig. 2.5 which

represents the main components of stand alone PV system.

Due to the variable nature of the energy source (solar radiation), one of the most

expensive aspects of a PV power system is the necessity to build in system autonomy.

Autonomy is required to provide reliable power during "worst case" situations, which

are usually periods of adverse weather, seasonally low radiation values or unpredicted

increased demand for power. The addition of autonomy could be accomplished by

oversizing the PV array and greatly enlarging the battery storage bank - generally the

two most costly system components. Substantial operating cost savings (compared to a

generator-set system without PV) are achieved through the greatly reduced need for

generator-set operation. An additional benefit of this approach is the added system

reliability provided by the incorporation of the back-up energy source.

Hybrid systems may contain more than one renewable power source. Adding a wind

turbine to a PV generator-set system is a common combination in areas with -high

wind energy potential [2.5].

2.5.2 Grid Connected PV Systems

A grid-connected photovoltaic system is an electrical power generating system that

uses a photovoltaic (PV) array as the primary source of electricity generation. It is

intended to operate synchronously and in parallel with the electric utility network. As

77
shown in Fig. 2.6, such systems may also include battery storage, other generating

sources, and may operate on site loads independent of the utility network during

outages [2.6]. Photovoltaic power systems have made a successful transition from

stand-alone to large grid-connected systems where the photovoltaic modules are

connected to the grid. There are a two ways in which the modules can be integrated

into the grid: connected photovoltaic arrays on the residential, commercial sites (i.e.

on rooftops) or, large arrays of photovoltaic modules that act as generating stations.

Fig. 2.6 Block diagram of a grid-connected PV system

Grid connected photovoltaic systems can vary in size but have common components

including: PV arrays, an inverter to convert the DC power to AC power, a controller

(i.e. protect the PV arrays from any voltage and frequency deviations), metering

equipment to monitor the power that is feed into the grid and vice versa, MPPT

(Maximum Power Point Tracker), module and other structures for mounting. Arrays

from residential and commercials can sell their excess power to the grid during the day

when more power than needed is obtained. When these sites need more power, (i.e. at

night) they can buy it from the grid. Thus, the grid acts as storage means for the PV

78
arrays during the day. Sometimes these PV arrays also have batteries and can choose

whether to store the excess power in the batteries or sell it to the grid.

In recent years, technological advances and price reductions have resulted in numerous

applications and increasing markets for grid-connected PV systems on residential and

commercial buildings. The primary component in grid-connected PV systems is the

inverter, or power-conditioning unit (PCU). The PCU converts the DC power

produced by the PV array into AC power consistent with the voltage and power

quality requirements of the utility grid. A bi-directional interface is made between the

PV system AC output circuits and the electric utility network, typically at an onsite

distribution panel or service entrance. This allows the AC power produced by the PV

system to either supply on-site electrical loads, or to back feed the grid when the PV

system output is greater than the on-site load demand. At night and during other

periods when the electrical loads are greater than the PV system output, the balance of

power required by the loads is received from the electric utility. When the utility grid

is down, these systems automatically shut down and disconnect from the grid. This

safety feature is required in all grid-connected PV systems to ensure that the PV

system will cease to operate and feed back into the utility grid when the grid is down

for service or repair.

2.6 Modeling of PV Systems

Photovoltaic (PV) cells are basically transducers that directly convert solar energy into

electrical energy. The physics behind such cells is similar to that of the basic p-n

junction [2.3, 2.4]. It follows that an appropriate model of the PV cell should be able

to predict its electrical characteristics under different irradiance and temperature

79
levels. Most of the widely used models of PV cells follow the equivalent circuit

approach where a PV cell is represented by an equivalent circuit that consists of one or

two diodes [2.7-2.12]. A simple and popular model of the PV cell is the one

representing it with an equivalent circuit consisting of a single diode [2.7]. In such a

model, the ohmic losses can be taken into consideration by including a series

resistance and/or a shunt resistance. When both the series and shunt resistances are

considered, the model of the PV cell, which is shown in Fig. 2.7 (a), requires

computation of five parameters in order to establish the current-voltage relationship

that characterizes the cell [2.8]. The number of parameters becomes four when only

the series resistance is taken into consideration

[2.9]. The four parameters model is shown in Fig. 2.7 (b). An attractive feature of the

aforementioned models is that most of the calculations only rely on the data provided

by the manufacturer. Another important feature is the fact that these models can be

used in representing a single cell, a module of connected cells, or even an array of

modules.

The single diode model is suitable for system-level designs. However, experiments

requiring high accuracy at the expense of complication can use more complicated

models such as the two diodes model [2.12]. In this model, an additional diode is

placed

80
 (a) (b)

Fig. 2.7 The equivalent circuit in single diode PV models. (a) Five parameters
model. (b) Four parameters model.

in parallel to that of the single diode model shown in Fig. 2.7 (a). To completely

establish the current-voltage characteristics of the cell, eight parameters need to be

calculated, hence the complication. A comprehensive review of the equivalent circuit

models can be found in [2.13].

An important issue to consider when modeling a PV cell is the cell temperature. The

cell temperature mainly depends on ambient and operating conditions. This

dependency is accounted for using what is called the thermal models of the PV cell

[2.8, 2.14]. A popular lumped thermal model has been proposed in [2.8] where the

heat losses are all lumped together in one coefficient. This thermal model, in

conjunction with the four parameters single diode model, is ideal for system level

design. Detail thermal models have also been proposed in the literature as in [2.14].

In this section, a practical PV cell is modeled and simulated. First, the four parameter

single diode PV model is discussed in section 2.6.1. A Simulink program based on this

model is developed in section 2.6.2. In section 2.6.3, the developed program is used in

the simulation of a practical PV cell and the results are shown to demonstrate its

functionality.

2.6.1 The Four-Parameter Single-Diode Equivalent Circuit Model

The equivalent circuit of a PV module according to this model is shown in Fig. 2.7 (b).

The model consists of a single diode, and the ohmic losses are represented by a series

81
 resistance. The expression governing the relationship between the load current and the

output voltage in the single-diode model is as follows:

U I Rs
I IL ID IL I0 exp 1 (0.12)

where IL , ID , I0 , and I are the light, dark, saturation and load currents, respectively, all

measured in (A); U is the output voltage (V); Rs is the series resistance (Ω); and is the

thermal voltage timing completion factor (V). For the I U relationship to be completely

defined, the quadruplet IL ,I R0, s , needs to be determined. For this reason, the

discussed single diode model is also called a four parameter model. The four

parameters depend on IL , temperature, and/or solar irradiance. This dependency will be

discussed in what to follow.

First, the calculation of the parameter is discussed. To calculate , the parameter

ref needs to be determined first as in [2.15]:

2U mp ref, Uoc ref,

I
ref Isc ref, Isc ref, mp ref, ln 1 IImp refsc ref,, (0.13)

where Ump ref, (V) and Imp ref, (A) are the maximum power point voltage and current,

respectively, at the reference condition; Uoc ref, (V) is the open circuit voltage of the

module at the reference condition; and Isc ref, (A) is the short circuit current of the

module at the reference condition. The values of Ump ref, , Imp ref, , Uoc ref, , and Isc ref, should

be provided by the manufacturer at the reference condition which is associated with a

82
specific reference irradiance level and a specific reference temperature. The parameter

depends on the temperature as follows:

Tc refT,c 273273ref
(0.14)

where Tc (°C) is the PV cell temperature, and Tc ref, (°C) is the reference temperature

which is usually taken to be 25 °C. After the calculation of , the series resistance Rs can

be easily found using the following expression [2.15]:

Rs ref ln 1 Imp refsc ref,, U mp ref, Uoc ref, (0.15)

I
mp ref,

However, the previous expression is an estimation of the value of Rs which in most

cases is provided by the manufacturer. The two remaining parameters are the currents

IL and I0 . The saturation current I0 is calculated using the following expression:

3 s
I0 I0,ref T c ref, 273273 exp eqgaprefN 1Tc refT,c 273273
(0.16)
Tc
where egap (J) is the band gap energy of the material which is 1.17 eV for silicon; N s is

the number of cells connected in series in the PV module; q (C) is the charge of an

electron; and I0,ref (A) is the saturation current at the reference condition and can be

calculated as follows [2.16]:

Uoc ref,
I0,ref IL ref, exp ref (0.17)

83
where IL ref, (A) is the light current at the reference condition, and it can be obtained

from the PV module data sheet supplied by the manufacturer. The last parameter to be

calculated is IL , and it can be found using:

IL IL ref, I ,SC Tc Tc ref, (0.18)

ref

where (W/m2) is the irradiance; ref (W/m2) is the reference irradiance; and I ,SC

(A/°C) is the temperature coefficient of the short-circuit current which is provided by

the manufacturer.

2.6.1.1 Thermal Model of the PV Module

The cell temperature Tc used in the calculation of the four parameters of the PV model

is not necessarily the temperature of the surrounding environment (ambient

temperature). However, the two are related along with the output voltage and the load

current in the thermal model the PV cell [2.16] which is given by the following

differential equation:

dTc kin PV, U I kloss Tc Ta (0.19)

CPV dt A
where CPV [J/(°C.m2)] is the overall heat capacity per unit area of the PV cell; kin PV, is

the transmittance-absorption product of PV cells; kloss [W/(°C.m2)] is the overall heat

loss coefficient; Ta (°C) is the ambient temperature; and A (m2) is the effective area of

the PV cell.

84
2.6.2 Simulink Implementation of the PV Model

In this section, a Simulink program is developed for the PV model based on (2.7)

(2.13). The block diagram for the Simulink implementation is shown in Fig. 2.8,

whereas a detailed block diagram of the PV module is shown in Fig. 2.9. The program

is conducted in two steps. The first step is to calculate the four parameters of the single

diode model with the irradiance, cell temperature, and the manufacturer data as inputs.

After these parameters are calculated, the program solves the I U relationship

numerically using the algebraic constraint block as shown in Fig. 2.9.

Fig. 2.8 Block diagram for the PV model implementation in Simulink.

85
 Fig. 2.9 Detailed Block diagram for the PV module of Fig. 2.8.

The manufacturer data are provided to the program by double-clicking the PV module

block diagram in Fig. 2.8. The data entry window is shown in Fig. 2.10 (a). Before

being processed by the program, simple initialization commands are carried out and

are shown in Fig. 2.10 (b). In Fig. 2.9, two Matlab function blocks are present. The

first called “Parameters Calc” is responsible for calculating the four parameters of the

single diode model based on the manufacturer data, irradiance, and cell temperature.

The code for this function is listed in Fig. 2.11 (a). The second function block called

“IV Calc” is responsible for the calculation of the current I in (2.7) for a given voltage

U and the knowledge of the model’s four parameters. The code for the “IV Calc”

function is listed in Fig. 2.11 (b).

86
 (a) (b)

Fig. 2.10 Dialogs related to the manufacturer data. (a) Data entry. (b) Initialization
code.

(a)

(b)
Fig. 2.11 Code listings for the function blocks used in the Simulink program.

87
 (a) Parameters Calc. (b) IV Calc.

2.6.3 Simulation of a PV module using the Simulink Program

The developed Simulink program will be used in this section to simulate the behavior

of a PV module whose data are reported in [2.16]. The module consists of 150 cells

that are connected in series. The data of this module are listed in Table 2.1. Two

studies will be conducted on the PV module using the Simulink program. In the first

study, the effect of the irradiance on the performance of the PV module is

investigated. This is done by simulating the module under irradiance values between

200-1000 W/m2 in 200 W/m2 steps while maintaining the cell temperature at a constant

level which is taken to be 25 °C in our case. The I U characteristic curves of the PV

module for the different values of irradiance are shown in Fig. 2.12. It can be observed

from the curves that the short circuit current Isc and the open circuit voltage Uoc of the

PV module increase with increasing irradiance levels. Accordingly, the maximum

power of the module increases.

For the different irradiance levels, the relationship between the output voltage and the

output power is shown in Fig. 2.13.

In the second study, the effect of the cell temperature on the performance of the PV

module is investigated. This is done by simulating the module with the temperature

values 0 °C, 25 °C, and 50 °C, while maintaining the irradiance at its reference level

of

1000 W/m2. The I U and P U characteristic curves of the simulated PV module for

Table 2.1 Parameters of the PV Module

Parameter Value

88
 ref 5.472 V

Rs 1.324 Ω

ref
1000 W/m2

Tc ref, 25 °C
I
IL ref, sc ref, 2.664 A

Ump ref, 70.731 V

Uoc ref, 87.72 V

Imp ref, 2.448 A

CPV 5×104 J/(°C.m2)

kin PV, 0.9

kloss 30 W/(°C.m2)

A 1.5 m2

I ,SC 0.00148 A/°C

the different values of temperature are shown in Figs. 2.14 and 2.15, respectively. It

can be observed that the maximum power and Uoc increase with decreasing

temperature.

However, decreasing the temperature leads to a slight decrease in Isc .

89
 I-V characteristics
3

1000 W/m 2

2.5

800 W/m 2
2

600 W/m 2
I (A)
1.5

400 W/m 2
1

200 W/m 2
0.5

0
0 10 20 30 40 50 60 70 80 90 100
V (V)

Fig. 2.12 I U curves for different irradiance levels at Tc Tc ref, .

P vs. V
200

180

160

140

120
P (W)
100 Increasing

80

60

40

20

0
0 10 20 30 40 50 60 70 80 90 100
V (V)

Fig. 2.13 P U curves for different irradiance levels at Tc Tc ref, .

90
 I-V characteristics
3

2.5

2 0 C

I (A) 25 C
1.5

50 C
1

0.5

0
0 10 20 30 40 50 60 70 80 90 100
V (V)

Fig. 2.14 I U curves for different temperature values with .

ref

P vs. V
200

180
0 C

160 25 C

140
50 C
120
P (W)
100

80

60

40

20

0
0 10 20 30 40 50 60 70 80 90 100
V (V)

Fig. 2.15 P U curves for different temperature values with .

ref

91
References

[2.1] Martin A. Green, Solar Cells Operating Principles, Technology and System
Applications, Published by the University of New South Wales, 1992.
[2.2] D. Yogi Goswami, Frank Kreith and Jan F. Kreider, Principles of Solar
Engineering, Taylor & Francis, USA, 2000.
[2.3] M. R. Patel, Wind and Solar Power Systems, CRC Press LLC, 1999.
[2.4] R. Messenger and J. Ventre, Photovoltaic Systems Engineering, CRC Press
LLC, 2000
[2.5] D. Hancen, H. Lars and H. Bindner, "Models for stand alone PV system", Riso
National Laboratory, Roskilde, Norway, December 2000.
[2.6] Bouna Ould Zeidane, "Contribution to the Study Of Grid Connected
Photovoltaic System", M.Sc, Université De Batna, 2003.
[2.7] Loferski J. J. (1972) An introduction to the physics of solar cells. In Solar Cells—
Outlook for Improved Efficiency, pp. 47–53, National Academy of Sciences–
Space Science Board, Washington, DC.
[2.8] Duffie, J.A., Beckman, W.A., 1991. Solar Engineering of Thermal Processes. John
Wiley & Sons Inc, New York.
[2.9] Rauschenbach H. S. (1980) Solar Cell Array Design Handbook: the Principles
and Technology of Photovoltaic Energy Conversion, Van Nostrand Reinhold,
New York.
[2.10] Roger, J.A., Maguin, C., 1982. Photovoltaic solar panels simulation including
dynamical thermal effects. Solar Energy, 29(3):245-256.
[2.11] Green, M. A. (1982). Solar cells: operating principles, technology, and system
applications. Englewood Cliffs, NJ, Prentice-Hall, Inc., 1982. 288 p., 1.
[2.12] Gergaud B, Multon B, Ben Ahmed H (2002) Analysis and experimental
validation of various photovoltaic system models. In: Proceedings of the 7th
international ELECTRIMACS’2002 congress, Montréal
[2.13] Rekioua, D., and Matagne, E. (2012). Optimization of photovoltaic power
systems. Springer.
[2.14] Mayer O. (1996). DASTPVPS Simulation Tool Manual, Universitat der
Bundeswehr Munchen, Neubiberg, Germany.
[2.15] T.U. Townsend, A Method for Estimating the Long-Term Performance of Direct-
Coupled Photovoltaic Systems, MS thesis, University Of Wisconsin – Madison,
1989.
[2.16] Øystein Ulleberg, Stand-Alone Power Systems for the Future: Optimal Design,
Operation & Control of Solar-Hydrogen Energy Systems, Ph.D. Dissertation,
Norwegian University of Science and Technology, Trondheim, 1998.

92
 Chapter 3 Improvement of PV Panel Performance Using
Switched Capacitor Converters

3.1 Improvement of Energy Extraction from Photovoltaic Panels with

Multilevel Output DC-DC Switched Capacitor Converters

3.1.1 Introduction

The total installed cost ($/W) and total cost of ownership ($/Wh) have been well-

studied as the key metrics controlling the grid penetration of solar power [3.1-3.4].

Among the factors impacting installed cost (per Watt) are power converter cost and

total (tracking x conversion) efficiency, both of which share strong relations to

converter and system complexity. A critical factor impacting the cost of ownership is

the lifetime of the power converter (and implied replacement costs). Cost-effective

solutions for solar energy extraction should address system cost and complexity,

conversion and tracking efficiencies and converter lifetime simultaneously.

The need for suitable tracking efficiency is normally addressed with a maximum

power point tracking (MPPT) algorithm embedded in the control of the converter or

inverter [3.5, 3.6]. In the important grid-tied case, 100 Hz power ripple at the panel

terminals negatively impacts the MPPT function, but this may be addressed by

augmenting the source with a large electrolytic capacitor [3.7-3.15]. However, the

limited lifetime of electrolytic capacitors contends directly with the long-life

characteristic of costeffective solar conversion. To reconcile this, [3.1] proposed the

"ripple port" inverter, which still directly interfaces the PV cell, but directs the 100 Hz

ripple power to a transformer coupled ripple port and away from the cell.

93
There is a growing need to implement per panel MPPT to contend with varying light

levels, temperatures, panel ages, etc. across physically widespread solar arrays [3.7-

3.15]. There are advantages of a DC-DC MIC + central DC-AC approach over a

DCAC MIC approach. These include the availability of a single DC bus for an entire

array and intermediate power ripple filtering, as well as added degrees of freedom for

MPPT control [3.7, 3.8, 3.16]. With per-panel MPPT, global tracking efficiency can

be significantly improved over simple series or parallel connections of those panels

(see Section 3.1.2.4), but the installation of per-panel converters impacts the important

cost metrics above. Converter lifetime and replacement costs become even more

critical with per panel conversion.

Many inductor based converters and inverters have been proposed as module

integrated converter (MIC) topologies, but they require magnetic components to be

either purchased per panel or to be integrated into the converter IC [3.7, 3.11].

Multilevel converters have been proposed as per panel DC-AC converters, but they

suffer from either 120 Hz power ripple at the panel terminals or the need for cost-

prohibitive and / or electrolytic energy storage [3.12-3.15].

94
Fig. 3.1 A DC linearized model with an N-panel PV string illustrates the
systemlevel approach. The ideal transformers model the function of the DC-DC
MICs.
3.1.1.1 System Overview

The system level approach that we are pursuing for this joint project is illustrated by

the DC linearized circuit model in Fig. 3.1. This system shares some key features with

other systems employing DC-DC MICs and a central inverter but differs in at least one

key way [3.7]. The DC-DC MICs typically operate with local autonomous MPPT

control. In the system proposed here, the responsibility of MPPT is shared among the

DC-DC modules and the central inverter. As a result, the required complexity of the

DC-DC MICs is simplified. Significantly, the system can be implemented with

switched capacitor multilevel DC-DC converters and a central ripple-port inverter. Per

panel magnetics are eliminated as are electrolytic capacitors. Any magnetics that are

required for the ripple port inverter need only be purchased once per string. The DC-

DC module conversion ratios are selectable, but discrete. A central question that is

addressed in this work concerns the tracking efficiency that is possible with this

system.

3.1.1.2 Switched Capacitor Benefits

Switched capacitor converters achieve current and voltage conversion without

magnetic energy storage. Fig. 3.2 shows the cost and volume per energy storage (J for

a sample of discrete capacitors and inductors suitable for power applications. These

data imply apriori expected cost and volume benefits of switched capacitor converters

when compared to inductor-based converters. A detailed discussion can be found in

[3.17].

95
Fig. 3.2 Specific cost and volume: Discrete inductors (10 µH- 1 mH/100 mA-1 A)
and capacitors (Ceramic and Film 1-10 µF/10-100 V) sampled from Digikey.
Energy was calculated as 1/2 CV2 or 1/2 LI2 for maximum rated voltages and
currents.

3.1.1.3 Total Efficiency

Total efficiency is central to the design and evaluation of the systems in this work.

Here we define total efficiency, η, as the product of tracking efficiency, η p, and

conversion efficiency, ηc:

p c (0.20)

Fig. 3.3 depicts a sample of reported tracking and conversion efficiencies in MPPT

algorithms and DC-DC MICs respectively. The two ranges are multiplied yielding a

third range corresponding to total efficiency η.

96
 Fig. 3.3: A literature survey of total energy extraction efficiency, DC-DC MICs:
[3.7, 3.8, 3.18–3.23] and MPPT algorithms: [3.6, 3.23–3.27]

3.1.2 Maximum Power Point Tracking

Maximum power point tracking in the system of Fig. 3.1 is simplified by the input

current control of the central inverter and the series connection of the MICs. The

selectable conversion ratios, Qi, allow the DC-DC modules to track local MPP’s as the

string current varies. The central inverter tracks the global MPP by adjusting its input

current.

The run-time global MPPT can be implemented by exploiting time-scale separation.

Here, we take the local MPPT control to operate fast, and the global MPPT control to

operate relatively slowly. Specifically, on the time-scale of local MPPT control, the

string current, Io, may be taken to be static or “quasi-static.” Because the maximum

power point of each panel is defined by a unique maximum power current, I mp,i, the

quasi-static string current naturally decouples MPPT control among the modules.

97
3.1.2.1 PV Model

Before discussing system performance further, we establish a model of the

photovoltaic panel and its parameters. The circuit model used here (Fig. 3.4) is

common in the literature, e.g. [3.16]. Given the parameters quoted in a typical

datasheet, Voc, Isc, and the maximum power voltage and current, V mp, and Imp, analysis

of the circuit in Fig. 3.4 yields (3.2)-(3.5).

Vdp Voc (0.21)

V
oc V mp
Rs (0.22)
Imp

Rp I Rsc sIVscoc (0.23)

Imp

Vfwd
I ph Imp (0.24)
Rp

where Vdp is the forward voltage of the diode, Dp, in the model. In Fig. 3.4, when Iin <
Imp, the diode, Dp, is forward-biased and it is reverse-biased otherwise. The resulting
panel voltages are:
Vin Vdp I Rins , Iin Imp

(0.25)
Vin R Ip ph Rs Iin Imp
Rp Iin ,

98
 Fig. 3.4: PV circuit model

and the panel power is simply:

Pin I Vin in (0.26)

For simplicity, the rest of this work assumes the following nominal datasheet values

adapted from a Mitsubishi PV-MF170EB4 [3.28]:

Imp 6.93A

V mp 24.6V
(0.27)
Isc 7.38A Voc
29V

3.1.2.2 Local Maximum Power Point Tracking

Local MPPT control consists of matching the string current, to the panel's own Imp,i's.
From Fig. 3.1, the ith panel current is

Iin i, Q Ii o (0.28)
Given a quasi-static string current, Io, the modules each choose a Qi to maximize their

panel power. This maximization step can be performed a number of ways. For

instance, the modules may estimate their Imp i, 's via short-circuit current measurements.

References [3.29]-[3.33] discuss MPPT control by this method. A perturb and observe

step may be necessary for good accuracy following the initial Imp guess. In our

99
 system, having discrete conversion ratios, a maximum of two additional observations

should be required to determine the actual optimal conversion ratio.

In the simulations that follow, the local algorithm for choosing conversion ratios was

implemented as follows. Given Imp i, either by the short-circuit method described

above or otherwise, the modules attempt to minimize the error Iin i, Imp i, . This

minimization is constrained according to the nonlinear behavior of the PV indicated in

Fig. 4. Combining (3.6) and (3.7), the panel power for the ith panel can be written:

Pin i, Iin i, Vdp i, Iin i2 , Rs,i Iin i, I mp,i

(0.29)

Iin i, Rp i, I ph,i Rs,i Rp,i Iin i2 , Iin i, I mp,i

Pin i,

I
Taking the derivative of (3.10) with respect to in i, yields
Pin i, Vdp i, 2Iin i, Rs,i Iin i, I mp,i
Iin i,

Pin i, Rp i, I ph,i 2Rs,i Rp,i Iin i, Iin i, I mp,i

Iin i, (0.30)
The term, 2R R I
s,i p,i in i, , in the derivative typically leads to a steep decrease in

panel power for Iin i, Imp i, . Absolute errors Iin i, Imp i, ; impact the panel power

less for Iin i, Imp,i . Accordingly, the algorithm adopted in this work attempts to

minimize the error Iin i, Imp i, with the following order of preference:

1)Iin i, I mp,i

2)Iin i, I mp,i

100
3)Iin i, I mp,i

In the examples presented here, the DC-DC modules each continuously attempt to

match the string current to their own MP currents according to the above algorithm.

Generally, the modules can choose from a set of integral conversion ratios [0,1...Q max].

The Q = 0 module configuration is important for good average tracking efficiency. It

is equivalent to the passthrough mode discussed in [3.7] and represents the option for a

panel to “sit out” when its maximum power is so low that including it in the string

would have a negative impact on the global MPP.

3.1.2.3 Global Maximum Power Point Tracking

The string inverter can track the global MPP by adjusting its input current. Fig. 3.5

depicts an example of the tracking efficiency achieved as Io is swept, while the DC-

DC modules adjust their conversion ratios. For this example, and for the rest of this

section, tracking efficiency is considered in an otherwise lossless system ( η c =

100%). Section 3.1.4 addresses the effect of converter efficiency on MPPT control.

The Io sweeps, like the one depicted in Fig. 3.5, may be performed on a scheduled

basis. Alternatively, Io may be varied continuously according to a particular runtime

MPPT algorithm. Section 3.1.4.5 presents a simulation of a likely input current-

controlled inverter.

101
Fig. 3.5 A single Io sweep with the following paramters: 3 panels, Qavail = [0,1,2,3,4],
Imp;vec = [6.898, 4.503, 4.878] A, ΔIo = 1 mA

3.1.2.4 Statistical Performance Evaluation

A statistical performance evaluation method was adopted to account for variations in

panel MPP’s. Monte Carlo simulations were performed by allowing MATLAB to

choose random (normalized) Imp i, 's for each panel. For each simulation, the string

current, Io, was swept as in Fig. 3.5 and the maximum efficiencies (tracking, converter,

and total) were recorded. Repeating this many times and averaging the results yielded

a prediction of average performance.

An example output plot is shown in Fig. 3.6. The plot in Fig. 3.6 reveals that tracking

efficiency can be very high for only a few panels. As panels are added, η p diminishes

to a limited extent. The local MPPT algorithm implemented impacts this behavior

significantly. For instance, if the order of preferences listed in Section 3.1.2.2 is

reversed, the tracking efficiency diminishes steadily as panels are added rather than

flattening as it does in Fig. 3.6. The Monte Carlo simulation results also show how

102
average tracking efficiency improves as the number of available levels increases. The

tracking efficiency predicted for a 3-panel, 5-level system is approximately 90%.

Increasing the number of available levels to 8 increases the predicted tracking

efficiency to 95%.

Fig. 6: Monte Carlo performance prediction: Qavail = [0 : 1 : Qmax], Monte Carlo

Length = 200, Io,sweep = [0.01 : 0.02 : 6.93] A
Finally, it should be noted that the Q max = 1 case (i.e. Q avail = [0; 1]) is somewhat

representative of a simple series string of panels with bypass diodes. The statistical

data predict roughly 65% average tracking efficiency while a 5-level MIC would

improve that efficiency to roughly 90%

3.1.2.5 Effect of spatial panel separation

In the above example, the panels are assumed to have a random and uncorrelated

distribution of MPP’s. Intuitively, this model becomes inappropriate as panels become

closely spaced. To model the effect of statistical correlation between MPP’s for panels

arranged in a non-infinite area, the randomly assigned panel MPP’s can be constrained

103
to a fraction of the full range. The simulation above was repeated having forced the

MPP’s to lie within 50% of the full range for each Monte Carlo iteration. The results

show universally higher average tracking efficiencies. For instance, the tracking

efficiency predicted for a 3-panel, 5-level system is approximately 95:5% and for a

3panel, 8-level system, 97:4%.

3.1.2.6 Non-integral level selections

So far, we have considered only switched capacitor multilevel converters having

integral, boosting sets of conversion ratios. There are many switched capacitor

topologies that can achieve rational and bucking conversion ratios as well. Such a

topology choice may be beneficial when considering upper bounds on DC bus

voltages or other practical issues. A more thorough investigation will be the subject of

future work.

3.1.3 Switched Capacitor Implementation

One particular realization of the switched capacitor MICs in Fig. 3.1 is the Marx

Multilevel converter. By forming series and parallel combinations of the the input

source and the switched capacitors, the 5-level Marx converter shown in Fig. 3.7 can

achieve conversion ratios Qavail = [0; 1; 2; 3; 4].

3.1.3.1 Efficient Switching Patterns

Switching cycles consist of a recharge phase, ϕ1, and an output phase, ϕ2. During ϕ1,

the switched capacitors are disconnected from the load and charged in parallel with the

source. During ϕ2, one of several series-parallel configurations of the switched

104
capacitors and input source is chosen to achieve the desired conversion ratio. Many

redundant switching configurations are possible. The switching configurations shown

in Fig. 3.8 were chosen for the 5-level Marx to minimize the conduction losses that

will be quantified shortly. Generally, the switching configurations were chosen to

minimize capacitor droop and the number of switches in the output current path, both

of which lead to loss and load regulation. Rules of thumb to minimize capacitor droop

include 1) utilize the input source to drive the output during ϕ 2 when possible and 2)

utilize all of the switched capacitors when driving the output, e.g. parallel-connect

redundant capacitors when possible.

Fig. 3.7 A 5-Level Marx converter

Fig. 3.8 Switching configurations

Switched capacitor circuits can achieve very high conversion efficiency by minimizing

the instantaneous current flow through their effective output resistance, R out,i. In a

105
DCDC switched capacitor circuit, the output is slowly-varying on the time-scale of

one switching period. These facts guide us to particular modes of operation. In

particular, efficient operation can be achieved when the same output phase (ϕ 2)

configuration is repeated every cycle. In contrast, modulation of the ϕ 2 configuration

on a per cycle basis, e.g. to achieve intermediate conversion ratios, would be ill-

advised as it would lead to continuously varying open circuit converter voltages

resulting in high instantaneous currents (high AC rms currents) through R out,i. This

observation leads directly to the constraint that the Marx Multilevel converter can

(efficiently) achieve a discrete set of conversion ratios.

3.1.3.2 Linear Modeling

A linear modeling approach was adapted from the work in [3.34]. This linear modeling

effort yielded quantitative support for the linear circuit models shown in Fig. 3.1

including the output resistances, Rout,i which represent both loss and load regulation in

the switched capacitor circuits [3.34].

3.1.3.3 Switching Speed Limit Definitions

According to [3.34], loss and load regulation mechanisms can be differentiated among

two switching speed limiting cases. In the slow-switching-limit (SSL), the switched

capacitors fully equilibrate yielding impulsive capacitor currents. In the fast

switchinglimit (FSL), the switched capacitors maintain fixed voltages while capacitor

currents during each switching state are constant [3.34]. The two switching speed

limits can be understood by considering the classic capacitor charging loss problem

depicted in Fig.

106
3.9. The total energy lost in charging the capacitor is the time-integral of I C(t)2R:

Etot Vs VC 0 2 RC e 2t RC/ (0.31)

2R
0

In the SSL, the exponential term is allowed to collapse to -1 and the energy lost

becomes

Etot SSL, C V C2 (0.32)

Fig. 3.9 The canonical circuit for studying the fundamental loss associated with
charging a capacitor.
independent of R, and in agreement with the classical result. In the FSL, (3.12) can be

viewed near t = 0 with the Taylor series approximation to the exponential term. This

leads to

Etot FSL, ( )t s VC 0 t
(0.33) V 2

i.e. the loss we would expect for two fixed voltages connected across the resistor.

Reference [3.34] shows how these two loss mechanisms yield asymptotic limits to the

output resistance with proportionalities

RSSL 1

107
 Cf
sw (0.34)
RFSL Rds on,

The method developed in [3.34] for computing the multipliers to quantify R SSL and

RFSL was adapted to the Marx Multilevel converter here. The results are summarized in

Tables 3.1 and 3.2 for Marx converters having between two and eight available levels.

Note that the multipliers in the tables need to be computed for each conversion ratio
(switching pattern) for each number of available levels (topology). Also note that R FSL
depends on the duty ratio between ϕ 1 and ϕ2, which was taken as D = 0.5 for all
switching patterns here. Given the asymptotic limits, the actual output resistance for
any combination of topology, C, fsw, and Rds;on is generally

Rout ! maxRFSL ,RSSL (0.35)

and the conduction loss per module is simply

Prloss I Ro2 out (0.36)

Table 3.1 RSSL Multipliers: (x1/Cfsw)

Levels Available: 2 3 4 5 6 7 8

Q=0 0 0 0 0 0 0 0

Q= 1 0 0 0 0 0 0 0
Q=2 - 1 1/2 1/3 1/4 1/5 1

Q=3 - - 2 3/2 1 5/6 2/3

Q=4 - - - 3 5/2 2 3/2

Q=5 - - - - 4 7/2 3

Q=6 - - - - - 5 9/2

Q=7 - - - - - - 6

108
 Table 3.2 RFSL Multipliers: (xRds,on)
Levels Available: 2 3 4 5 6 7 8

Q=0 2 4 6 8 10 12 14
Q=1 2 4 6 8 10 12 14

Q=2 - 8 10 12.4 8.2 17.6 32.4

Q=3 - - 26 24 38 48.4 50.8

Q=4 - - - 64 90 100 100

Q=5 - - - - 130 180 206

- - - - 232 307
Q=6 -
Q=7 - - - - - - 378

3.1.3.4 Switching Loss

The linearized model above captures loss due to output current conduction. When

evaluating the design in Section 3.1.4, it will also be important to include switching

loss, a loss mechanism not explicitly contained in the linearized circuit model of Fig.

3.1. The switching loss for any active switch (one that changes state between the two

switching phases) can be quantified by considering the circuit shown in Fig. 3.10. All

MOSFETs in the Marx converter reside in at least one loop consisting only of one or

two other MOSFETs and a switched capacitor. In the Marx converter, the switched

capacitor, C, in Fig. 3.10 will nominally exhibit a voltage equal to the panel voltage,

109
Vin, because it is recharged to that potential each cycle. The total switching loss was

estimated in terms of typical data sheet values using [3.35] for N active devices as

Pswloss N Q V g g 1Qloss Vin Q Vrr in f sw (0.37)

2

Examining the switching patterns shown in Fig. 3.8, one can extract the following

pattern generalizing the number of active switches according to conversion ratio:

N 1, Q 0
(0.38)
N 3Q 2, Q 0

Fig. 3.10 Switching loss evaluation in the Marx converter for active MOSFETs

3.1.3.5 Inherent Features

Inherent to the topology of the Marx converter are a few interesting features that may

add significant value to a solar power system. As mentioned previously, the Marx

converter has a natural pass-through feature, replicating the function of bypass diodes

and also the pass-through mode presented in [3.7]. The ability to disconnect each

module from the load may be beneficial when implementing safety disconnect

features. Reference [3.36] discusses the need for a disconnect in the event of a fire to

110
prevent electrocution hazards that would otherwise result from the high voltage string

output. This disconnect feature may also be particularly beneficial in implementing an

antiislanding mode. A good discussion of anti-islanding control and solutions for solar

power systems can be found in [3.37].

The run-time local MPPT algorithm described above can be designed to automatically

prevent under-voltage conditions at the panel output. Because the DC-DC modules

I I
continuously choose Qi to closely match in i, to mp,i , they automatically adjust to over-

current conditions, choosing Qi = 0 in the limiting case. This feature is advantageous

when the local control circuitry is powered by the panel itself.

3.1.3.6 Gate Drive

The gate drive for the Marx converter needs to operate with a continuous floating gate

drive voltage. The converter itself does not guarantee a periodic charging path to

recharge a bootstrap capacitor. A level shift circuit is also required to translate

groundreferenced logic signals to the gate drive output. The recommended topology

(not necessarily the specific parts), adapted from [3.38], is shown in Fig. 3.11.

111
 Fig. 3.11 The recommended gate drive adapted from IR AN-978 [3.38].

The 555 timer IC, 1N4148 diodes and floating capacitor form a charge pump circuit.

The 100 kΩ resistor between the 555 timer GND and Panel GND and the 15 V zener

allows the timer IC to float 15 V below the source of the driven MOSFET.

The low power version of the 555 timer IC (ICM755) is needed in this circuit to

achieve low power dissipation in the part itself and to achieve sufficient quiescent

current despite the 100 k resistor to ground. The charge pump drives the VB node to

twice its supply voltage referenced to its own floating GND leading to a 15 V floating

drive referenced to the MOSFET source. This voltage can be adjusted by choosing the

voltage of the Zener diode. The “HV Level Shift” could be a commercial high side

driver IC such as the IR2125. However, the high voltage rating of such a part would

be underutilized for a typical implementation of the system in this work. Therefore, a

more costeffective gate drive would include a custom level shift circuit.

112
3.1.4 Design Example

A 3-panel 510 W system was designed and simulated in SPICE and in MATLAB.

Among the key topological considerations for implementing a practical Marx DC-DC

MIC is the need for a power diode in series with the output of each module. This diode

is required to block current from conducting backwards through the body diode of the

upper MOSFET in the output stage during ϕ1. In order to alleviate any need to

synchronize switching action among modules, a local output non-electrolytic capacitor

was placed across each module to create a local DC bus.

3.1.4.1 Number of Levels

The number of levels was chosen using the same Monte Carlo prediction methods

described in Section 3.1.2.4. Having enumerated loss mechanisms, total efficiency was

used to determine performance. To choose an appropriate number of levels, an

unoptimized but lossy system was simulated using nominal circuit parameters and

MOSFET device characteristics. The predicted performance is plotted in Fig. 3.12.

The data show diminishing returns in total efficiency beyond 5 levels. Therefore a 5-

level Marx converter was chosen as the MIC.

3.1.4.2 MOSFET Choice and Switching Frequency Optimization

Having chosen a reasonable value for the non-electrolytic (metal film) switched

capacitors in the 5-level marx, the choice of MOSFET and switching frequency were

optimized together. It is significant that all MOSFETs in the Marx converter reside in

loops containing a switched capacitor and other MOSFETs only. MOSFET drain-

113
source voltages are therefore upper bound by the maximum panel voltage.

Accordingly, it is

Fig. 3.12 Unoptimized system performance prediction: Qavail = [0 : 1 : Qmax], Monte

Carlo Length = 400, Io;sweep = [0.01 : 0.02 : 6.93] A, C = 12.5 µF, fsw = 250 kHz,
Rdson = 10 m, Qg = 10 nC, Qoss = 5 nC, Qrr = 25 nC, Vg = 15 V, Voc = 29 V, Vmp =
24.6V, Isc = 7.38 A, Imp = 6.93 A, Distribution Compression = 50%

particularly advantageous to choose a panel whose open-circuit voltage is just below a

standard value for Vdss. Over-sizing the MOSFET beyond the required Vds rating

would lead to unneeded switching or conduction loss and a suboptimal design. A

number of likely MOSFETs were identified having V dss = 30 V for the panel open-

circuit voltage of 29 V. The likely MOSFETs were chosen based on their on-

resistance, Rds;on, and gate capacitance, Cg. With the losses derived in (3.17) and (3.18),

the performance was plotted for each MOSFET across switching frequency and

conversion ratios. Fig. 3.13 shows such a plot for the selected MOSFET. A maximum

average converter efficiency (across all conversion ratios) of > 98% was predicted at a

switching frequency of 360 kHz. Note that the gate drive voltage was decreased from

114
15 V in the unoptimized system to 10 V in the optimized system. Adjusting the gate

drive voltage trades off conduction loss (on-resistance) for switching loss.

Fig. 3.13 MOSFET and switching frequency evaluation at peak power:

IRF8721,
C = 12.5 µF, Vmp = 24.6 V, Imp = 6.93 A, MP = 170 W, Vg = 10 V

3.1.4.3 Power Diode

The power diode was chosen primarily to support the peak output current and to block

the peak reverse voltage safely. Secondly, it was chosen for low capacitance, forward

voltage, and ESR. Having added output diodes to the implemented system, the

additional losses can be estimated as follows:

Vfwd i, ln s o 1n q ESRdiode oI I (0.39)

Pdiode Io " Vfwd i, f swC j i, Vrr i2, (0.40)

I kT

115
 i

where Vrr is the reverse voltage during ϕ 1 and Cj is the junction capacitance of the

diode. This expression can be used to improve the accuracy of the Monte Carlo

performance predictions. The power diode chosen for this example was the Motorola

MBR20100C Shottky [3.39]

3.1.4.4 Simulated Prototype

The optimized system was simulated using SPICE and MATLAB. The performance of

this system was predicted with Monte Carlo methods having incorporated the losses

derived in (3.17), (3.18) and (3.21). The results are shown in Table 3.3. A summary of

circuit elements selected for the simulated prototype is shown in Table 3.4.

Table 3.3 Simulated statistical performance: 5-level, 3 Panel optimized system:

Monte Carlo Length = 100, Distribution Compression = 50%, ΔIo = 1 mA, Diode Loss
= [on]

Efficiency symbol simulated result

tracking Ʉ୮ 95.43%

conversion Ʉୡ 97.56%

total Ʉ 93.10%

Table 3.4 Circuit component summary

Component Part No. / Value Note

Switched
Capacitors, 12.5 µF Metal Film

1||4.7 || 6.8 µF
Output Capacitor

116
 Panel Capacitor 25 µF 12.5 || 12.5 µF

MOSFET IRF8721

Output Diode MBR20100C

It is important to realize that the central inverter cannot track panel power,

corresponding to ƞp, directly. Instead it tracks its input power, corresponding to ƞ.

Having incorporated the loss mechanisms from Section 3.1.3 and in (3.21), this

observation was accounted for in simulation by allowing the inverter to choose the I o

that maximized its input power. Tracking efficiency was recorded for comparison to

total efficiency. A single experiment was performed in simulation to validate the linear

modeling effort and loss calculations above. A fixed set of conversion ratios and

MPPs was chosen for the three panels. Tracking, conversion, and total efficiencies

were plotted for a single Io sweep. Fig. 3.14 compares the results for calculated data

based on Section 3.1.3 and (3.21), a SPICE simulation of the linearized model and a

SPICE simulation of the MOSFET system. The difference in ƞc between the linearized

model

117
Fig. 3.14 Model Validation: Single Io sweep, 3 sources, Q = [0, 2, 4], Imp;vec = [0.007
3.465 6.93] A, C = 12.5 µF, f sw = 360 kHz, MOSFET: IRF8721, Vg = 10 V,
deadtime = 100 ns, Rg = 4 Ω
and the other two data sets represents switching and output diode loss. Errors between

the calculated model and FET simulation are likely due to estimation errors in

computing diode and switching losses. Note that in the plots of Fig. 3.14, the

maximum in total efficiency lines up closely with the maximum in tracking efficiency.

Time domain waveforms from the simulated system are shown in Fig. 3.15. Fig.

3.15(a) shows a zoom-in of the capacitor currents. The shape of those currents

indicates operation between the slow and fast switching limits defined in Section

3.1.3.3.

This result is a natural outcome of the MOSFET choice and switching frequency

optimization step above. Fig. 3.15(b) shows panel input currents during a step change

in the load current from 90% to 100% of the predicted maximum power current.

118
In this example, Panel 1 is bypassed (Q 1 = 0) because its MPP is quite low; I in1 = 0 in

the plots. The other two panels initially settle close to their respective I mp;i’s - Panel 2

exhibits half of the photovoltaic current that Panel 3 does. When the load current steps

to its maximum power value, Iin2 and Iin3 settle on their respective Imp;i’s.

Fig. 315 Time-domain waveforms.

3.1.4.5 DC AC Dynamics

A linearized model of the central input current-controlled inverter is shown in Fig.

3.16.

The closed-loop transfer functions of particular interest can be derived from that
circuit.

They are:

v ref 1 A s( )

j s( ) e s( )

119
 A s( ) Gc ( )s Fm sLe MRe 2 (DR ) 1sC
(0.43)

F s( )HRsense (0.44)

T s( ) AF (0.45)

Fig. 3.16 A linearized model of an input current-controlled inverter front-end.

The linear model parameters, e, j, M(D), L e, and Re were chosen for a 500 W

buckderived inverter topology [40]. Fig. 3.17 shows step responses of the closed-loop

transfer functions in (3.22) and (3.23). They show relatively fast settling times in the

input current upon step transients in the input voltage (corresponding to the string DC

bus voltage) and the reference voltage (corresponding to the control for the sweepable

input current). The lower plot also indicates a significant attenuation of the input

current response to changes in the input voltage. This attenuation is largely dependent

on the low-frequency magnitude of the loop gain T(s) as indicated by (3.22).

120
 Fig. 3.17: Inverter closed-loop step responses

3.2 Improved Transient Response Control Strategy and Design

Considerations for Switched- Capacitor (SC) Energy Buffer

Architectures

3.2.1 Introduction

A power conversion interface between a DC and a single-phase AC system requires an

energy buffer to store the instantaneous power difference between the constant power

of the DC port and the time-varying power of the AC port. Traditionally, this energy

buffer is implemented with large electrolytic capacitors. As the system reaches

121
periodic steady state with unity power factor, the instantaneous power difference

manifests itself as a voltage ripple on the energy buffering capacitor at double-line

frequency. However, the single capacitor approach is known to achieve poor energy

utilization [3.41].

Energy utilization is defined as the ratio of the energy used to buffer the instantaneous

power difference to the maximum stored energy on the capacitor. Energy utilization

for single capacitor energy buffers with respect to the peak-to-peak ripple ratio can be

derived as:

2r
Eutil 2
(0.46)
1 r 0.5r
where r is the prescribed peak-to-peak ripple ratio. For instance, in a system with 10%

peak-to-peak ripple ratio, the single capacitor energy buffer implementation has an

energy utilization of less than 20%. In other words, the capacitor has to store more

than 5 times the energy as actually needed.

Many alternative techniques have been proposed to manage the double-frequency

energy flow [3.41-3.45]. In particular, the recently introduced switched-capacitor (SC)

energy buffer architectures [3.46, 3.47] can achieve higher energy utilization and

lower

122
 Fig. 3.18 General architecture of the SC energy buffer.

voltage ripple. Fig. 3.18 shows the general architecture of the proposed SC energy

buffer. The SC energy buffer consists of two banks of capacitors shown as one

possible example in Fig. 3.18: "backbone" capacitors and "supporting" capacitors. The

configuration will be described as y-z, where y is the number of capacitors in the

backbone bank and z is the number of capacitors in the supporting bank.

The backbone capacitor bank contains capacitors that withstand large voltage

variations during the ripple cycle, where the voltage variations are typically much

greater than the prescribed peak-to-peak ripple allowance. In order to bring the bus

voltage ripple within bound, the supporting capacitor bank is switched so that the

voltages of the supporting capacitors are either added to or subtracted from the voltage

of the backbone capacitor bank. The switching pattern is defined such that the

resulting bus voltage satisfies the ripple specification. The supporting capacitors have

to withstand a much smaller voltage variation during the ripple cycle. Specifically, in

123
this two-bank energy buffer architecture, the voltage variations on the supporting

capacitors are limited to one-half the specified peak-to-peak bus ripple magnitude if

the supporting capacitors and backbone capacitors are equally sized.

Using this technique with a peak-to-peak ripple ratio of 10%, energy utilization can be

improved to >70% with one backbone capacitor and >80% with three backbone

capacitors. Moreover, this technique enables the use of capacitors with smaller

capacitance and lower voltage ratings, thereby making it possible to replace limited-

life electrolytic capacitors with ceramic or film capacitors. Practical uses of this

technique require control schemes that can produce acceptable transient responses to

timevarying power levels. Section 3.2.2 examines different control schemes and

exposes undesirable behavior under certain operating conditions. The second part of

this section proposes a two-step control scheme while taking a comprehensive view of

the design space, exploring tradeoffs between circuit topology and control. Topology

selection and switching configurations are discussed in Section 3.2.3. Control strategy

requirements and tradeoffs are examined in in Section 3.2.4. Finally, simulation results

of the proposed control strategy in two different circuit topologies are presented in

Section

3.2.5.

3.2.2 Control Algorithms

Different control schemes have been proposed for the SC energy buffer shown in Fig.

3.18 [3.46, 3.47]. Two approaches are reviewed briefly here, neither of which is

necessarily satisfactory during transient operation. These two approaches illustrate

different control design philosophies with regard to the bus voltage ripple relative to

124
the overall level of power being processed. In the first case, discussed below in the

context of a PFC utility interface for providing power to a load, the controller offers a

fixed peak-to-peak bus ripple, ΔV r , pp , for all but very low power levels during steady-

state operation. In the second case, discussed in the context of a grid-tie inverter, the

controller varies the steady-state bus ripple with the power level injected to the grid. In

both cases, power delivery changes can result in large bus voltage transients.

3.2.2.1 Bus-Voltage Monitoring, Finite State Machine Control

We first illustrate a control problem in the case of a SC energy buffer inside a PFC

utility interface. In this example, discussed in [3.47], the controller directly monitors

the bus voltage and triggers finite-state-machine state transitions when the bus voltage

is about to exceed pre-defined bounds. The switching pattern associated with each

state is defined so that an increase in state number would boost the bus voltage up by

ΔVr , pp when the bus voltage dips below the lower trigger threshold, and a decrease in

state number would drop the bus voltage down by ΔV r , pp when the bus voltage rises

above the upper trigger threshold.

Because the supporting capacitor voltages are not individually monitored, state

transitions do not guarantee the desired boost or drop on the bus voltage, e.g., as

shown in [3.47]. Also, the state machine is unaware of the power level and is not reset

or “recentered” between ripple cycles, so power transients may cause the state to

saturate at either the state associated with the lowest or the highest apparent energy.

During this state saturation, the SC energy buffer no longer has any available state to

contain the ripple in the saturation direction. Finally, because the controller attempts to

125
maintain the bus voltage within constant DC boundaries at all times, a transient

response to a new

Fig. 3.19 Transient bus voltage response of a 2-6 SC energy buffer in a PFC due
to a 30% load power step. The energy buffer uses bipolar switching configuration
and is controlled by the bus voltage monitoring, finite state machine controller in
[3.46]. The state variable saturates at the terminal states, 1 and 24, during the
power transient, causing unacceptable ripple shootthrough.

steady-state power level can lead to extreme bus voltage transients as the controller

will attempt to maintain the DC boundaries until it is driven into state saturation.

To investigate such undesirable behaviors, a SPICE simulation is performed using

LTSpice from Linear Technology. A LT1249 active power factor controller is selected

for the simulation because the model is readily available in the bundled component

library. The simulated testbench circuit is derived from the typical application

example in the datasheet [3.48] with the output filter capacitor replaced by the 2-6 SC

energy buffer presented in [3.47]. In addition, the simulation model also incorporates

the controller implemented with a 24-state finite state machine and an "artificial

126
feedback voltage" reported in [3.47]. The design specifications include a nominal

output voltage of 320V and a 20% peak-to-peak ripple ratio. The simulation results are

shown in Fig.

3.19, where the bus voltage exhibits unacceptable over- and undershoots when the

state machine state saturates at states 1 and 24 in response to 30% load power level

transients. Note that the artificial feedback voltage does not faithfully reproduce the

over- and under-voltage conditions. The extreme overshoots from the shortcomings of

the controller are amplified by two additional factors. The capacitances of the

capacitors in the energy buffer are greatly reduced under the assumption of proper

ripple reduction. Moreover, the capacitors are linked in series, which further

diminishes the effective capacitance seen on the bus.

3.2.2.2 Supporting Capacitor Monitoring, Timing Interval Control

A similar control problem can be illustrated considering Fig. 3.18 in its inverter

configuration. In this case, discussed in [3.46], the individual supporting capacitor

voltages are monitored while giving up the task of controlling the backbone capacitor

voltage to the energy-balance controller of the inverter. The control logic pre-

computes the charge and discharge intervals for each supporting capacitor relative on

the phase of the ripple cycle and enables these intervals when the capacitor voltages

are within their reference minima and maxima [3.46]. The reference voltages scale

linearly with power level and the ripple is reduced by a fixed ratio. Therefore, the

resulting bus voltage behavior is very similar to that of a single capacitor

implementation – the backbone capacitor experiences the natural transient and settling

127
behaviors from the energybalance controller, and the supporting capacitors are used to

keep the ripple voltage within the prescribed limits.

However, this controller makes inefficient use of the supporting capacitor bank – all

capacitors in the supporting bank are used regardless of power level. As a result, the

supporting capacitor voltage references must be adjusted significantly in response to

power variations. Since the voltage on capacitors cannot change instantaneously, the

supporting capacitors will need time to be charged or discharged to the new reference

levels. This introduces a few cycles where the supporting capacitors experience large

imbalance in their charge and discharge times. In the extreme case, the supporting

capacitors may not be used in either the charge or the discharge cycle at all, thus

exposing the bus to the full-swing ripple from the backbone capacitor with reduced

capacitance during the corresponding half cycle.

A SPICE simulation is again used to demonstrate the potential problems with this

control strategy. The simulated testbench circuit is implemented using the feedforward

energy-balance controlled solar inverter demonstrated in [3.46] along with a 1-8 SC

energy buffer. The nine supporting capacitors are monitored and managed by the

controller with pre-computed switch timings discussed above, and the backbone

capacitor is controlled by the feedforward energy-balance controller of the solar

inverter. The design specifications include a nominal output voltage of 250V and a

10% peak-to-peak ripple at maximum power. As shown in Fig. 3.20, the bus voltage

experiences an unacceptable undershoot when the supporting capacitor voltages

references are dramatically increased in response to 30% power level transients. The

second subplot in Fig. 3.20 shows the nonparticipation of the supporting capacitors

128
during their discharge half-cycles, resulting in the lack of buffering during the

discharge half cycle.

3.2.3 SC Energy Buffer Design Considerations

There are many tradeoffs to be considered in designing an SC energy buffer. A basis

for making these tradeoffs is developed in this section. In principle, energy utilization

can

Fig. 3.20 Transient bus voltage response of a 1-8 SC energy buffer in a solar
inverter due to a 30% input power step. The energy buffer uses the unipolar
switching configuration and is controlled by the supporting capacitor monitor,
timing interval controller. The discharge is disabled in order to charge the
supporting capacitors up to the new reference values, exposing the full-swing
backbone capacitor ripple.

be increased arbitrarily at the expense of switching frequency and buffer complexity.

Desirable transient performance implies new control requirements that also impact SC

129
buffer design. We consider these tradeoffs in the context of two general SC buffer

architectures, unipolar and bipolar switching configurations shown in Fig. 3.21.

A first consideration in designing the energy buffer is energy utilization when the

design goal is to reduce the overall amount of physical capacitance in the system.

Equation (3.27) summarizes the energy utilization for a non-switching, single

capacitance buffer. The energy utilization equation can be generalized for the SC case

shown in Fig. 3.18 by taking the sum of ΔE , the change in energy stored, divided by

the

Figure 3.21 General 1-z architecture of SC energy buffer. (a) Implementation

with ground-referenced switches only for unipolar switching configuration. (b)
Implementation with four additional switches to achieve bipolar switching
configuration.

130
sum of Emax , the maximum energy stored, of all the capacitors in the energy buffer .

This is shown in (3.28).

y z

"Ebackbone( )j " Esupport ( )j

Eutil yj1 iz1
(0.47)

"Emax,backbone( )j "Emax,support ( )j j 1
i1

The variables in (3.28) depend on not only the nominal bus voltage, the specified

ripple ratio and the selected capacitor size, but also the switching configuration. Thus,

the two cases shown in Fig. 3.21 illustrate a tradeoff between topology and switching

complexity versus capacitor utilization. Note that Fig. 3.21 illustrates the two cases

with a single backbone capacitor, i.e., y = 1 in each case, although more backbone

capacitors could be employed with arbitrary y.

3.2.3.1 Capacitor Configurations

For illustration, the energy utilization of an SC energy buffer with bipolar switching

configuration versus different numbers of backbone and supporting capacitors for

three different ripple ratios is shown in Fig. 3.22. Two important conclusions can be

drawn from these plots. First, for each ripple ratio and number of backbone capacitors

used, there exists an optimal number of supporting capacitor which maximizes the

energy utilization of the overall energy buffer. Secondly, the energy utilization can be

improved with diminishing return by introducing more backbone capacitors.

131
However, the number of backbone capacitors cannot be increased indefinitely. The

switching frequency of the SC energy buffer is directly proportional to the number of

capacitors in the energy buffer. In particular, the switching frequency can be

approximated as

f f
sw !2 grid pyz 1 (0.48)

where p = 2 for unipolar switching schemes and p = 4 for bipolar switching schemes.

Clearly, increasing the number of capacitors would unavoidably increase the incurred

switching loss. Also, excessive number of capacitors would cause the SC buffer

switching frequency to approach that of the PFC or inverter controllers, consequently

causing undesirable interactions between the two control loops.

In order to guarantee time-scale separation between the low frequency energy buffer

control and high-frequency PFC or inverter control, the number of capacitors must be

limited. When designing a

switching converter, the

switching frequency is

expected to (a)

132
 (b)

(c)

Fig. 3.22 Overall energy utilization of the SC energy buffer as a function of the
capacitor configuration. These numbers are computed for SC energy buffers with
equally sized backbone and supporting capacitors using the bipolar switching
configuration. Recall that y denotes the number of backbone capacitors and z
denotes the number of supporting capacitors.

be high with respect to the natural frequency of the energy storage elements. This

extends to the case of a SC energy buffer. While any specific case requires a control

loop and stability analysis, a similar rule-of-thumb to keeping the natural time constant

in the canonical models long compared to the switching period, e.g. 10 times the

switching period, is to have the SC buffer switching at below 1/10 the frequency of the

133
interfacing switching converter. As illustrated in Fig. 3.18, high frequency switching

converters can be found on either side of the SC energy buffer.

For example, assuming the switching frequency of the high frequency loop is on the

order of a hundred kilohertz, average switching frequency of the energy buffer control

might be constrained to be less than approximately ten kilohertz. In other words, the

relationship in (3.30) must hold.

10kHz
pyz #1 (0.49)
f
2 grid

This establishes an upper bound on the number of capacitors that can be incorporated

in these SC energy buffers. Referring back to Fig. 3.22, the unfeasible combinations of

capacitor configurations are greyed out. As shown, the achievable improvement in

energy utilization is limited, albeit still significant, as this becomes a constrained

optimization problem. For peak-to-peak ripple ratios of 2%, 5%, and 10%, the optimal

achievable energy utilizations are realized with only one or two backbone capacitors.

In a SC energy buffer, the bus voltage is no longer an accurate measure of the energy

stored in the energy buffer. Therefore, when integrating with conventional power-

factor correction controllers or energy-balance inverter controllers, the bus voltage

cannot be directly used as the feedback voltage. Reference [3.47], for example, uses an

artificial feedback voltage to ensure compatibility with existing hardware. However,

such an artificial feedback voltage is not guaranteed to be sinusoidal and may not

reliably detect under- and over-voltage conditions as shown previously.

By implementing the backbone capacitor bank with only one capacitor, a voltage

feedback signal is available at the single backbone capacitor for interfacing with

134
conventional power-factor correction controllers or energy-balance inverter

controllers. Because there is a single path in the backbone capacitor bank through

which the energy buffering current must flow, the single backbone capacitor voltage

can be treated as an AC-scaled version of the single electrolytic capacitor voltage in

traditional energy buffers.

Energy utilization is still high with a single backbone capacitor. Specifically, in the

case of 10% peak-to-peak ripple ratio, using a single backbone capacitor reduces the

achievable energy utilization from 77.9% to 71.2%, still a sizable improvement from

18.1%. In the cases of 5% and 2% peak-to-peak ripple ratios, the optimal energy

utilizations remain unchanged. Also, this simplification enables the exclusive use of

ground-referenced switches in unipolar switching configurations.

3.2.3.2 Switching Topology Tradeoffs

We therefore focus on the 1-z architecture shown in Fig. 3.21, where we define N = z

+ 1 as the total number of capacitors in the SC energy buffer. The backbone capacitor

is denoted as C0, and the supporting capacitors are denoted as C 1 through CN-1. Two

types of switching configurations can be explored: unipolar and bipolar. In unipolar

switching, supporting capacitor voltages are added to the backbone capacitor voltage

when it is too low, but are never subtracted. With equally sized capacitors, the

resulting peak-to-peak bus voltage ripple with respect to the total number of capacitors

is

Vr pp unipolar, , N2
%
1 $o %C VP % C (0.50)

135
 $
where P is the power level, o is the angular frequency of the grid, C is the capacitance

of all capacitors in the SC energy buffer, and V C is the nominal voltage of the grid. If

the backbone capacitor voltage is regulated by energy balance control, i.e., to achieve

constant mean squared voltage, using the unipolar switching configuration will result

in a variable mean bus voltage. Specifically, the mean bus voltage will increase with

increasing power level, but will always be above the regulated mean voltage of the

backbone capacitor. For this reason, the unipolar switching configuration is unsuitable

for PFC applications with constant output voltage requirements. However, it is

compatible with solar inverters where the bus voltage must remain sufficiently high in

order to maintain control of the grid. In addition, because the mean bus voltage is

positively correlated to the power level, it ensures fast response time in hysteresis

current controlled inverters when the output current amplitude is increased. Finally,

the one-sided switching configuration also has the added benefit of being able to

utilize ground-referenced switches only. By rearranging the supporting capacitor bank

and the backbone capacitor as shown in Fig. 3.21 (a), the unipolar SC energy buffer

avoids high-side gate drives.

In the bipolar switching configuration, four additional switches are added in order to

invert the polarity of the supporting capacitor voltages during parts of the ripple cycle.

This enables ripple reduction with a constant mean bus voltage. Supporting capacitor

voltages are added to the backbone capacitor voltage when it is too low and are

subtracted from the backbone capacitor voltage when it is too high. As such, the

bipolar switching configuration is compatible with power-factor correction

applications without an additional dc-dc converter at the output. Moreover, the bipolar

136
switching configuration uses the supporting capacitors more efficiently; it achieves a

peak-to-peak voltage ripple of

Vr pp bipolar, , N1 %$o %C VP % C
(0.51)

approximately twice as effective, in terms of ripple reduction capability versus number

of capacitor added, as the unipolar switching configuration. The ripple adavantage

requires four extra switches and high-side gate drives, which contribute to additional

switching losses.

The steady-state maximum supporting capacitor voltages under maximum power

rating for both switching configurations are outlined here to supplement energy

utilization calculations and to facilitate capacitor selections.

V max,unipolar ( )i 1%$o %C Vmax% C (0.52)

N
i 1 P
V max,bipolar ( )i % max (0.53)
i 1 P

2N $o %C V% C

for i = {1,2, …., N-1}. For the backbone capacitor, the maximum capacitor voltage is

the same for both switching configurations and can be calculated as

V max(0) VC12%$oP%C Vmax% C (0.54)

137
In the following section, control strategies for both switching configurations are

presented.

3.2.4 Two-Step Control Strategy

A controller capable of handling power level transients must not prescribe strict DC

voltage boundaries constraints on the bus voltage. Instead it should allow the DC level

of the bus voltage to undergo natural settling while maintaining the AC ripple

magnitude within specification around the DC level. This enables the controller to

evenly distribute the charge buffering to the supporting capacitors instead of leaving

the terminal-state capacitors to absorb an unusual large amount of leftover charges.

Also, the controller must effectively reset its state from ripple cycle to ripple cycle in

order to guarantee the availability of reserve buffering states in the event of power

transients. Finally, the controller must intelligently manage the supporting capacitors

so they can remain effective in reducing the ripple magnitude at all time. This

translates to maintaining the reference voltage levels of the supporting capacitors

relatively constant regardless of power level.

These requirements can be satisfied by adopting a two-step control strategy: capacitor

participation optimization and switch timing determination. The controller first

determines the optimal number of capacitors to use in buffering the bus voltage, and

then compute the switch timings for the allocated supporting capacitors to maximally

reduce the bus voltage ripple. In a 1-z SC energy buffer configuration, the single

backbone capacitor voltage is used as the feedback node to either a PFC or an inverter

controller. Thus, the SC energy buffer controller discussed here passes the regulation

of the backbone capacitor voltage to an external interfacing controller.

138
Two design examples will be presented to better illustrate the operation and the

effectiveness of the proposed control strategy. The specification for the design

examples is a 500W inverter with a 250V nominal bus voltage and a 10% peak-to-

peak ripple

Fig. 3.23 Sampling points and control variables, v c(i) and, vd(i) in relation to the
ripple cycle and the control ramps for (a) unipolar switching configuration and
(b) bipolar switching configuration

ratio. For maximum energy utilization, a 1-8 SC configuration is chosen for the

unipolar switching scheme. For the bipolar switching scheme, a 1-4 SC configuration

is chosen for comparable switching complexity and ripple reduction power.

3.2.4.1 Capacitor Participation Optimization

In order to optimize the supporting capacitor participation, the controller samples the

current power level and calculates the minimum number of capacitors required to keep

the voltage ripple within the specification. The sampling frequency is twice the line

139
frequency for the unipolar switching configuration and four times the line frequency

for the bipolar switching configuration.

The sampling points with respect to the ripple cycle are illustrated in Fig. 3.23. Note

that the minimum required number should have a lower bound at 1 because the

backbone capacitor is always used, and can be derived by inverting the ripple

magnitude

(3.31) and (3.32) for the two different switching topologies. Equation (3.36) shows the

solution for the unipolar switching configuration and (3.37) shows the solution for the

bipolar switching configuration. Note that P[n] is the sampled power level during the

current ripple cycle.

Nunipolar[ ]n max ceil $ 1,1 (0.55)

o% %

%C VC V r pp,

,1
Nbipolar[ ]n max ceil P n[ ] (0.56)
2 [ ]P n

$o % % %C VC V r pp,

By only using the minimum required number of capacitors, the controller ensures that

there is a sufficient number of capacitors in reserve, ready to kick in during a sudden

power level increase. In addition, relatively constant energy storage in the supporting

capacitors is maintained over a wide range of power levels. Consequently, the system

is able to respond to large power transients by adjusting the number of capacitors used,

140
rather than drastically changing the energy stored on all the supporting capacitors. Fig.

3.24 illustrates the supporting capacitor voltages and the expected ripple size across all

possible power levels in the 1-8 unipolar SC energy buffer design example. The

number of switching events is reduced as the power level decreases, which improves

the overall system efficiency.

3.2.4.2 Switch Timing Determination

Given the number of capacitors to use, the controller proceeds to compute the switch

timings for the capacitors based on the current power level. That is, the charge and

discharge cycle durations are adjusted for each supporting capacitor based on the

current sample of its voltage and its respective reference values.

141
 Fig. 3.24 Expected ripple magnitude and the supporting capacitor voltages as a
function of power level for the 1-8 unipolar design example.

Since the charging and discharging of the capacitors by the double-line frequency

energy flow are inherently nonlinear with respect to time, a nonlinear element is

inserted into the control loop to enable the use of simple linear function in the rest of

the controller. The nonlinear element takes form of a control ramp on which the

switching event is triggered. For the unipolar switching configuration, the control

ramp is a double-line frequency sine wave phase-locked to the grid. In addition, the

unipolar control ramp is assumed to be normalized with unit peak-to-peak amplitude

and ramps from 0V to 1V.

142
For the bipolar switching configuration, the ripple cycle can be further broken up into

two sub-cycles. There is the additive sub-cycle where the supporting capacitor

voltages are added to the bus voltage, and the subtractive sub-cycle where the

supporting capacitor voltages are subtracted from the bus voltage. Thus, the same

cycle duration computation needs to be performed twice as often as in the unipolar

case. The control ramp function for the bipolar switching configuration then must be

periodic at four times the line frequency. Specifically, the bipolar control ramp is a

rectified and inverted version of the unipolar control ramp and ramps from 0V to

0.5V. The two control ramp signals in relation to their respective control voltages and

sampling points are shown in Fig. 3.23.

Because the control ramps are assumed to be normalized, the control equations will

also be defined in a power-independent fashion. All sampled values are normalized to

the full-swing ripple magnitude on the backbone capacitor. The normalizing function

is defined as

v[ ]n
v[ ]n (0.57)
$ %
P n[ ]/ o %C V C

where v[n] is the sampled supporting capacitor voltage.

Based on the normalized sampled supporting capacitor voltages, the allowable

discharge and charge durations for each capacitor are calculated from (3.39) and

(3.40),

disch( )i max min v i [ ]n Dx [ ]n , Dx [ ]n , Dx [ ]n (0.58)

(0.59)

143
 charg( )i max min v i [ ]nDi x[ ]n2 , Dx1[ ]n , Dxk[ ]n

i 1 k

Where i=[1,2…N-1] denotes the supporting capacitor index, 1/D x[n] is the normalized

step in voltage between the supporting capacitors, and k&[0,1] determines the

minimum duration. The variable x in D x [n] denotes the switching configuration. The

discrete step size definitions differ in the two switching configurations and are shown

in (3.41) and (3.42).

1 1
(0.60)
Dunipolar[ ]n Nunipolar[ ] 1n

1 1
(0.61)
Dbipolar[ ]n 2Nbipolar[ ]n

The minimum duration defined by k determines the tradeoff between transient ripple

size and settling time. If k is very close to zero, the controller may allow the capacitor

voltages to reach their new reference values quicker by imposing a large imbalance

between their charge and discharge cycles. However, larger imbalances between the

charge and discharge cycles increase exposure of the bus voltage to the ripples of the

backbone capacitor, resulting in larger transient ripple. If k is very close to one, the

controller will maintain ripple buffer throughout more of the ripple cycle. But the

limited imbalance between the charge and discharge cycles results in longer settling

times. Note that by managing the capacitor participation based on power level, the

144
reference voltages for the supporting capacitors are kept fairly constant. Therefore, k

can be set very close to one for adequate buffering without the risk of unreasonably

long settling times.

Having computed the allowable charge and discharge durations for each supporting

capacitor, the actual control voltages can be calculated by a cumulative sum. More

specifically, the individual charge and discharge control trigger levels are

N x [ ] 1n

vd ( )i " disch(m) (0.62)

mi

Fig. 3.25 Proposed two-level SC energy buffer controller block diagram, where V o
denotes the backbone capacitor voltage V i[n], for i=[1,2,..N-1] denotes the
sampled supporting capacitor voltage Vc, and Vd correspond to the charge and
discharge control signals respectively.

N x [ ] 1n

vc ( )i " charg(m) (0.63)

mi

When Nx[n]-1 < i, the control voltage is set to zero, which means that supporting

capacitor i is not being used in the current ripple cycle. Furthermore, higher-indexed

switches have precedence over lower-indexed switches. That is, if v c(1), vc(2), … , vc

(M) > vramp , switches 1, 2, … , M-1 are all disabled, and only switch M is turned on.

145
The complete two-step controller block diagram is shown in Fig. 3.25.

3.2.4.3 Distortion and Phase Error

In the previous section, the control ramps are assumed to be perfectly sinusoidal, or

rectified sinusoidal, with zero phase error. Practical phase-locked loops may not

guarantee zero steady-state phase error. If a phase error persists between the control

ramp and the actual ripple cycle, systematic errors would be introduced to the

steadystate voltages of all supporting capacitors, which would result in an increased

overall bus voltage ripple. Additionally, the grid voltage may not be perfectly

sinusoidal and the ripple voltage may exhibit distortions. Distortion from the assumed

sinusoidal profile would introduce unsystematic imbalances in the charge and

discharge of the supporting capacitors, which again causes the overall bus voltage

ripple to increase. Therefore, the generated phase-locked signal cannot always be

used. Instead, the control ramps can be derived from the backbone capacitor voltage.

By passing the AC component of the backbone capacitor voltage through a clamped

capacitor circuit, a unipolar control ramp signal from 0V to the peak-to-peak ripple

magnitude can be extracted. Similarly, the bipolar control ramp can be created by

inverse rectifying the AC component of the backbone capacitor voltage, then

processing the resulting signal with a clamped capacitor circuit. This yields a bipolar

ramp signal from 0V to the peak ripple amplitude. Alternatively, both control ramp

signals can be produced digitally after sampling the backbone capacitor voltage.

Generating the ramp functions directly from the backbone capacitor voltage

guarantees zero distortion and phase error between the control signals and the actual

ripple cycle. Furthermore, normalization of the sampled signals may not be required

146
because the normalization factor is the inverse of the peak-to-peak ripple amplitude on

the backbone capacitor. In practice, implementing control logic with the large voltages

may not be feasible. Therefore, resistive dividers can be employed as long as the

divider ratio is consistent between the control ramp generation and the supporting

capacitor sampling.

3.2.4.4 Pre-charge Circuit Requirement

It is not necessary to have a pre-charge circuit used in [3.47] when using the control

strategy described in the previous sections. By adjusting the switch timings, the

controller automatically introduces imbalances between the allowable charge and

discharge durations of the supporting capacitors so the capacitor voltages reach their

reference.

This is a tradeoff. The pre-charge circuit can facilitate the process of charging the

supporting capacitors to their reference levels at startup, which allows the system to

reach steady-state operation faster. Secondly, the pre-charge circuit can assist in

maintaining the charges on unused capacitors. The proposed controller only controls

charge and discharge duration on the active supporting capacitors in the ripple cycle; it

has no control over the nonparticipating capacitors in reserve. Thus, having a

precharge circuit adds an extra layer of security to ensure that the capacitors in reserve

remain ready in the event of a power level increase. Finally, by using a pre-charge

circuit to set up all the capacitors to known states initially, the SC energy buffer can in

principle be operated without a requirement to monitor the voltage on every

supporting capacitor in the buffer.

147
3.2.4.5 Over- and Under-voltage Protection

Aside from the overvoltage protection circuitry commonly found in PFC and inverter

controllers, the SC energy buffer controller can incorporate an additional layer of

protection to guard against large transients between sampling periods. Switching

duration computations are performed at the beginning of each sampling period. If the

transient between sampling periods is large enough, the computed and ideal switch

timings may differ significantly, resulting in over- or under-buffering conditions.

“Over-buffering” occurs when the actual ripple magnitude is significantly smaller than

the expectation of the controller. When such an event occurs, the boost and drop in the

bus voltage from switching the supporting capacitors will be greater than what is

actually needed. Similarly, “under-buffering” occurs when the actual ripple magnitude

is significantly larger than the expectation of the controller. Consequently, the boost

and drop in the bus voltage from switching the supporting capacitors will be smaller

than the required values. Both over- and under-buffering conditions result in larger

than expected ripple.

Such undesirable conditions can be avoided by introducing feedforward compensation,

i.e., a forced resampling triggered on over- and undervoltage thresholds. Once the bus

voltage exceeds the defined thresholds, the controller resamples the current power

level and the supporting capacitor voltages to recompute the number of active

capacitors required and recalculate the switch timings. In over-buffering conditions,

the recomputed number of active capacitors would be decreased, whereas in

underbuffering conditions, the recomputed number of active capacitors would be

increased.

148
3.2.5 Simulation Results

The unipolar 1-8 SC energy buffer and the bipolar 1-4 SC energy buffer design

examples have been successfully implemented and simulated in SPICE with a 500W

inverter. The system is implemented with control ramps generated from the backbone

capacitor voltage to avoid distortion and phase errors. In addition, the minimum

duration constant k is set to 0.9 and a pre-charge circuit is configured to manage the

voltages of supporting capacitors in reserve. The steady-state bus voltage ripple and

the backbone capacitor feedback voltage are shown in Fig. 3.26. The simulated result

matches the analytical solution quite well. The external inverter control manages the

backbone voltage and holds it to 250V. The peak-to-peak ripple is set to 10% by

inverting (3.31) and (3.32) and solving for the required capacitance.

149
Fig. 3.26 Steady-state bus voltage waveforms of the (a) 1-9 SC energy buffer with
unipolar switching experiencing increasing power level and (b) 1-4 bipolar SC
energy buffer with bipolar switching experiencing decreasing power level. In (a),
the power level increases from 96W to 480W with +48W step size every 50ms. In
(b), the power level decreases from 480W to 96W with a -96W step size every
50ms.

The bus voltage in the unipolar switching energy buffer exhibits a power-dependent

mean as discussed in Section 3.2.3, and remains well above the grid voltage to retain

control. As the power level increases, more supporting capacitors become involved in

ripple buffering, as demonstrated by the capacitor activities in the subplot of Fig.

3.26(a). Conversely, the bus voltage in the bipolar switching energy buffer has a

150
constant mean over the all power levels as shown in Fig. 3.26(b). With decreasing

power level, the supporting capacitors sequentially become inactive, leaving only the

backbone capacitor to buffer the small power ripple. In a sampled system, the

worstcase behavior occurs if a large transient occurs immediately after sampling has

taken place.

Thus, this is the case chosen for the transient response characterization. Positive and

negative 30% steps in input power level are introduced to the inverter with the bipolar

1-4 SC energy buffer. As shown in Fig. 3.27, the positive step in power causes an

underbuffering condition until the bus voltage crosses the upper threshold. Note that

the over and undervoltage thresholds are defined to be 1.5 times the ripple

specification, i.e.

15% peak-to-peak from 250V, and shown in Fig. 3.27 as dotted lines. At this point, the

Fig. 3.27 Transient bus voltage response of the example bipolar 1-4 SC energy
buffer in a solar inverter due to 30% input power step. The power steps from
480W to 336W at 50ms and back to 480W at 100ms. The second supporting

151
capacitor voltage is shown to deviate from its reference value shortly after 100ms,
but the two-step controller brings it back to its reference level in less than 2 ripple
cycles.
controller immediately resamples and recomputes the switch timings to pull the bus

voltage back within bounds. Even though the transient may cause some supporting

capacitor voltages,V2 in this particular example, to deviate from their reference values,

the two-step controller is able to bring the system back to steady-state in just a few

cycles, without any unacceptably large transient ripple.

3.3 Conclusion

Widespread grid penetration of PV will rely on the reduction of capital cost and total

cost of ownership for solar power systems. It is critical that these factors guide the

design of photovoltaic power circuits and system architectures. This work has

presented a full system approach utilizing switched capacitor multilevel DC-DC

converters. Substantial cost reductions may be possible by providing per panel MPPT

without the need for per panel magnetics. Coupling the DC-DC modules with a ripple

port inverter eliminates the need for electrolytic capacitors, enabling long-life

operation.

Topics of ongoing research include investigation of MPPT algorithms and related

system level tradeoffs for control of the central inverter. There exist necessary

tradeoffs among switching frequency, converter efficiency, and global tracking

efficiency (Io step size) when considering the dynamics and runtime MPPT

approaches for the full system. Not addressed in this work was the dual case of

paralleled strings of panels. This will also be a topic of ongoing investigation.

152
 Switched-capacitor energy buffers have been shown to achieve much better energy

utilization than their single electrolytic counterparts. However, overshooting and the

possibility of losing control to the grid are major concerns. The proposed control

strategy can potentially minimize the possibility of such undesirable behaviors by

maintaining an appropriate number of supporting capacitors in reserve to guard against

sudden transients in power level.

Two SC energy buffers – 1-8 with unipolar switching and 1-4 with bipolar switching –

have been examined in a 500W inverter. The simulated models show excellent

agreement with the calculated results. Furthermore, the system is able to maintain a

minimum bus voltage of 250V and limit the peak-to-peak ripple to 10% under

steadystate operation . It is also shown that the new control strategy can successfully

maintain the ripple specification under significant power level transients.

References

[3.1] P. Krein and R. Balog, “Cost-effective hundred-year life for single-phase inverters
and rectifiers in solar and led lighting applications based on minimum capacitance
requirements and a ripple power port,” in Applied Power Electronics Conference
and Exposition, 2009. APEC 2009. Twenty-Fourth Annual IEEE, feb 2009, pp.
620 –625.
[3.2] P. T. Krein, “Power electronics needs and performance analysis for achieving grid
parity solar energy costs,” power point presentation.
[3.3] R. W. Erickson, “Future of power electronics for photovoltaics,” power point
presentation.
[3.4] J. K. Mapel, “Organic photovoltaics and concentrators,” Ph.D. dissertation,
Massachusetts Institute of Technology,, 2006.
[3.5] T. Esram, J. Kimball, P. Krein, P. Chapman, and P. Midya, “Dynamic maximum
power point tracking of photovoltaic arrays using ripple correlation control,”
Power Electronics, IEEE Transactions on, vol. 21, no. 5, pp. 1282 –1291, sept
2006.
[3.6] R. Faranda, S. Leva, and V. Maugeri, “Mppt techniques for pv systems: Energetic
and cost comparison,” in Power and Energy Society General Meeting -
Conversion and Delivery of Electrical Energy in the 21st Century, 2008 IEEE,
july 2008, pp. 1 –6.

153
[3.7] L. Linares, R. Erickson, S. MacAlpine, and M. Brandemuehl, “Improved energy
capture in series string photovoltaics via smart distributed power electronics,” in
Applied Power Electronics Conference and Exposition, 2009. APEC 2009.
TwentyFourth Annual IEEE, feb 2009, pp. 904 –910.
[3.8] G. Walker and P. Sernia, “Cascaded dc-dc converter connection of photovoltaic
modules,” Power Electronics, IEEE Transactions on, vol. 19, no. 4, pp.1130 –
1139, july 2004.
[3.9] H. Patel and V. Agarwal, “Mppt scheme for a pv-fed single-phase single-stage
gridconnected inverter operating in ccm with only one current sensor,”Energy
Conversion, IEEE Transactions on, vol. 24, no. 1, pp. 256 –263, march 2009.
[3.10] B. Sahan, A. Vergara, N. Henze, A. Engler, and P. Zacharias, “A single-stage
pv module integrated converter based on a low-power current-source inverter,”
Industrial Electronics, IEEE Transactions on, vol. 55, no. 7, pp. 2602 –2609,
july 2008.
[3.11] W. Xiao, N. Ozog, and W. Dunford, “Topology study of photovoltaic interface
for maximum power point tracking,” Industrial Electronics, IEEE
Transactions on, vol. 54, no. 3, pp. 1696 –1704, june 2007.
[3.12] F. Filho, Y. Cao, and L. Tolbert, “11-level cascaded h-bridge grid-tied inverter
interface with solar panels,” in Applied Power Electronics Conference and
Exposition (APEC), 2010 Twenty-Fifth Annual IEEE, feb 2010, pp. 968 –972.
[3.13] S. Kouro, B. Wu, A. Moya, E. Villanueva, P. Correa, and J. Rodriguez,
“Control of a cascaded h-bridge multilevel converter for grid connection of
photovoltaic systems,” in Industrial Electronics, 2009. IECON ’09. 35th
Annual Conference
of IEEE, nov 2009, pp. 3976 –3982.
[3.14] E. Ozdemir, S. Ozdemir, and L. Tolbert, “Fundamental-frequency-modulated
sixlevel diode-clamped multilevel inverter for three-phase stand-alone
photovoltaic system,” Industrial Electronics, IEEE Transactions on, vol. 56,
no. 11, pp. 4407 –4415, nov 2009.
[3.15] M. Mousa, M. Ahmed, and M. Orabi, “New converter circuitry for pv
applications using multilevel converters,” in Telecommunications Energy
Conference, 2009. INTELEC 2009. 31st International, oct 2009, pp. 1 –6.
[3.16] P. Krein and Kim, “Pv converter module configurations for mppt operation,”
PECI, 2010.
[3.17] M. Seeman, V. Ng, H.-P. Le, M. John, E. Alon, and S. Sanders, “A comparative
analysis of switched-capacitor and inductor-based dc-dc conversion
technologies,” in Control and Modeling for Power Electronics (COMPEL),
2010 IEEE 12th Workshop on, june 2010, pp. 1 –7.
[3.18] E. Roman, R. Alonso, P. Ibanez, S. Elorduizapatarietxe, and D. Goitia,
“Intelligent pv module for grid-connected pv systems,” Industrial Electronics,
IEEE Transactions on, vol. 53, no. 4, pp. 1066 –1073, june 2006.
[3.19] G. Walker and J. Pierce, “Photovoltaic dc-dc module integrated converter for
novel cascaded and bypass grid connection topologies 8212; design and
optimisation,” in Power Electronics Specialists Conference, 2006. PESC ’06.
37th IEEE, june 2006, pp. 1 – 7.

154
[3.20] W.-Y. Choi, S.-M. Kim, S.-J. Park, K.-H. Kim, and Y.-C. Lim, “High step-up
dcdc converter with high efficiency for photovoltaic module integrated
converter systems,” in Telecommunications Energy Conference, 2009.
INTELEC 2009. 31st International, oct 2009, pp. 1 –4.
[3.21] B. Yuan, X. Yang and D. Li, “Analysis and design of a high efficiency current
fed multi-resonant converter for high step-up power conversion in renewable
energy harvesting,” in Power Electronics for Distributed Generation Systems
(PEDG), 2010 2nd IEEE International Symposium on, june 2010, pp. 138 –
143.
[3.22] J. Enslin, M. Wolf, D. Snyman, and W. Swiegers, “Integrated photovoltaic
maximum power point tracking converter,” Industrial Electronics, IEEE
Transactions on, vol. 44, no. 6, pp. 769 –773, Dec. 1997.
[3.23] J.-M. Kwon, B.-H. Kwon, and K.-H. Nam, “Three-phase photovoltaic system
with three-level boosting mppt control,” Power Electronics, IEEE
Transactions on, vol. 23, no. 5, pp. 2319 –2327, sept 2008.
[3.24] D. Hohm and M. Ropp, “Comparative study of maximum power point tracking
algorithms using an experimental, programmable, maximum power point
tracking test bed,” in Photovoltaic Specialists Conference, 2000. Conference
Record of the Twenty-Eighth IEEE, Jan. 2000, pp. 1699 –1702.
[3.25] Y. Bo, Y. Xu, and L. Donghao, “Application of power compensating concept
for high efficiency maximum power point tracking in grid-connected
photovoltaic system,” in Power Electronics Conference (IPEC), 2010
International, june 2010, pp. 938 –941.
[3.26] C. Sullivan and M. Powers, “A high-efficiency maximum power point tracker
for photovoltaic arrays in a solar-powered race vehicle,” in Power Electronics
Specialists Conference, 1993. PESC ’93 Record., 24th Annual IEEE, Jun.
1993, pp. 574 –580.
[3.27] J. Lopez-Seguel, S. Seleme, P. Donoso-Garcia, L. Morais, P. Cortizo, and M.
Mendes, “Comparison of mppt approaches in autonomous photovoltaic energy
supply system using dsp,” in Industrial Technology (ICIT), 2010 IEEE
International Conference on, march 2010, pp. 1149 –1154.
[3.28] M. E. Corp., “Mitsubishi electric photovoltaic module pv-mf170eb4 170wp,”
Datasheet, June 2006.
[3.29] J. Martynaitis, “Discussion of ”theoretical and experimental analyses of
photovoltaic systems with voltage and current-based maximum power point
tracking”,” Energy Conversion, IEEE Transactions on, vol. 19, no. 3, p. 652,
sept. 2004.
[3.30] H. Matsuo and F. Kurokawa, “New solar cell power supply system using a
boost type bidirectinal dc-dc converter,” Industrial Electronics, IEEE
Transactions on, vol. IE-31, no. 1, pp. 51 –55, feb. 1984.
[3.31] T. Noguchi, S. Togashi, and R. Nakamoto, “Short-current pulse based adaptive
maximum-power-point tracking for photovoltaic power generation system,” in
Industrial Electronics, 2000. ISIE 2000. Proceedings of the 2000 IEEE
International Symposium on, vol. 1, jan 2000, pp. 157 –162 vol.1.
[3.32] C. Yu-yun and M. Yong-kui, “Constant current-based maximum-power-point
tracking for photovoltaic power systems,” in Control and Decision Conference,
2009. CCDC ’09. Chinese, june 2009, pp. 3422 –3425.

155
[3.33] S. Yuvarajan and J. Shoeb, “A fast and accurate maximum power point tracker
for pv systems,” in Applied Power Electronics Conference and Exposition,
2008. APEC 2008. Twenty-Third Annual IEEE, feb. 2008, pp. 167 –172.
[3.34] M. Seeman and S. Sanders, “Analysis and optimization of switched-capacitor
dc x2013;dc converters,” Power Electronics, IEEE Transactions on, vol. 23,
no. 2, pp. 841 –851, march 2008.
[3.35] I. Rectifier, “Irl3714z hexfet power mosfet,” Datasheet, 2004.
[3.36] R. Ram, “Power electronics,” ARPA-E Pre Summit Seminar, March 2010.
[3.37] P. Krein and T. Esram, “Admittance control for anti-islanding protection in
modular grid-connected inverters,” in Control and Modeling for Power
Electronics (COMPEL), 2010 IEEE 12th Workshop on, june 2010, pp. 1 –5.
[3.38] I. Rectifier, “An-978 hv floating mos-gate driver ics,” Application Note.
[3.39] M. S. Inc., “Switchmode power rectifiers mbr20100c,” Datasheet, 1996.
[3.40] J. Kassakian, M. Schlecht, and G. Verghese, Principles of Power Electronics.
Adison-Wesley Publishing Co., 1992
[3.41] Krein, P.T.; Balog, R.S.; , "Cost-Effective Hundred-Year Life for Single-Phase
Inverters and Rectifiers in Solar and LED Lighting Applications Based on
Minimum Capacitance Requirements and a Ripple Power Port," Applied Power
Electronics Conference and Exposition, 2009. APEC 2009. Twenty-Fourth
Annual IEEE , vol., no., pp.620-625, 15-19 Feb. 2009
[3.42] Kyritsis, A.C.; Papanikolaou, N.P.; Tatakis, E.C.; , "Enhanced Current
Pulsation Smoothing Parallel Active Filter for single stage grid-connected AC-
PV modules," Power Electronics and Motion Control Conference, 2008.
EPEPEMC 2008. 13th , vol., no., pp.1287-1292, 1-3 Sept. 2008
[3.43] Shimizu, T.; Jin, Y.; Kimura, G.; , "DC ripple current reduction on a single-
phase PWM voltage-source rectifier," Industry Applications, IEEE
Transactions on , vol.36, no.5, pp.1419-1429, Sep/Oct 2000
[3.44] Pierquet, B.J.; Perreault, D.J.; , "A single-phase photovoltaic inverter topology
with a series-connected power buffer," Energy Conversion Congress and
Exposition (ECCE), 2010 IEEE , vol., no., pp.2811-2818, 12-16 Sept. 2010.
[3.45] Bush, C.R.; Wang, B.; , "A single-phase current source solar inverter with
reduced-size DC link," Energy Conversion Congress and Exposition, 2009.
ECCE 2009. IEEE , vol., no., pp.54-59, 20-24 Sept. 2009
[3.46] Chang, A.H.; Cooley, J.J.; Leeb, S.B.; , "A systems approach to photovoltaic
energy extraction," Applied Power Electronics Conference and Exposition
(APEC), 2012 Twenty-Seventh Annual IEEE , vol., no., pp.59-70, 5-9 Feb.
2012
[3.47] Minjie Chen; Afridi, K.K.; Perreault, D.J.; , "Stacked switched capacitor energy
buffer architecture," Applied Power Electronics Conference and Exposition
(APEC), 2012 Twenty-Seventh Annual IEEE , vol., no., pp.1404-1413, 5-9 Feb.
2012.
[3.48] Linear Technology, “LT1249 – Power Factor Controller,” Datasheet

156
 Equation Chapter (Next) Section 1Chapter 4
Maximum Power Point Tracking of PV Arrays in Kuwait

4.1 Introduction

One of the available technologies to generate electricity from renewable resources is to

use solar cells to convert the energy delivered by the solar irradiance into electricity.

Currently, PV energy generation has a great commercial and academic interest. Recent

studies indicate that in the medium to longer term PV generation may become

technically and commercially worldwide feasible to be implemented for electricity

generation. The main purpose of this report is to study the performance of PV arrays

under varying weather conditions in Kuwait. For this study a data base of hourly solar

radiation and temperature are collected for a period of six years.

The grid connected PV system converts sunlight directly into ac electricity to supply

local loads and inject the excess energy to the public grid. The main purpose of the

system is to reduce the electrical energy imported from the electric network. Fig. 4.1

shows the block diagram of the basic components for a grid-connection of PV system.

The dc output of the PV array is converted into ac and injected into the grid through an

interfacing inverter. The controller of this system implements all the main regulation

and protection functions such as:- Maximum Power Point Tracking (MPPT),

protection relay and detection of islanding operation.

157
 Fig. 4.1 Basic components of grid connected PV array

4.2 Maximum Power Point

Under constant irradiance and cell temperature, the operating point of a PV array is

determined by the intersection of the Ipv-Vpv characteristic and the load line as shown in

Fig. 4.2. The load characteristic is defined by a straight line with a gradient given by

ILoad/VLoad. The system operating point moves along the Ipv-Vpv characteristic curve of

the PV panel from B to A as the load resistance increase from zero to infinity. Position

C is the maximum power operating point. If the load resistance is too high, the

operating points would be in the CA regions. If the load resistance is too low, the

operating points would be in the CB regions. Therefore, the maximum power point

can be obtained by matching the load resistance to the PV array characteristics.

158
 Fig. 4.2 PV characteristic curve and load line

4.3 Effect of Irradiance and Cell Temperature

The effect of irradiance and cell temperature on Ipv-Vpv characteristic curve is shown in

Figs. 4.3 and 4.4, respectively. Fig. 4.3 indicates that the maximum power output is

directly proportional to the irradiance. On the other side, Fig. 4.4 shows that the

maximum output power from the array decreases as the temperature increases.

As illustrated, the Ipv-Vpv characteristic of a PV cell varies with the irradiance and

operating temperature of the cell. The temperature of the panel, or more precisely the

diode junction temperature, also affects the voltage-current curve of the panel. This

effect can be taken into a consideration by using temperature coefficients. The typical

values of the coefficient for silicon and germanium diodes are:

dV
2.3.....2.1mV / C (1.1) dT

It should be noted that these coefficients are negative. This means the higher the

temperature, the lower the open-circuit voltage. The voltage-dependency can be

159
considered linear. Thereby, the voltage vs. temperature variation could be represented

as

Fig. 4.3 Effect of irradiance on the maximum power at constant cell temperature

160
Fig. 4.4 Effect of temperature on the cell characteristic at constant irradiance
percentage coefficients. Hence, these coefficients provide the percentage voltage
variations with the temperature changes. These coefficients stay the same regardless of the
amount of panels connected in series. The voltage as a function of temperature can thus be
found according to:
V T V( ) o[1 ( T T Ko) V ] (1.2)

where
dV
KV 100% (1.3)
dT

The value KV can be found on most PV-panel data-sheets [4.1, 4.2].

A previous study was carried by T. Nordmann and L. Clavadetscher to define the

effect of solar radiation on the panel temperature in different kinds of installations

[4.1, 4.3]. According to this study, the temperature of the panel may rise between 20 oC

and 55oC above ambient per kW/m2 of insolation. This means that the panel

temperature will be between 45 oC and 80 oC on a 25 oC day at 1000 W/m2 solar

irradiation.

As a result, the environmental factors clearly have a large effect on solar panel

behavior. This is why panel specifications are given in standardized conditions. Two

sets of environmental parameters have been standardized in IEC 60904-3. Standard

test conditions (STC) are defined so that panel temperature is 25 OC, insolation is 1000

W/m2, and ambient temperature is not defined. Normal operating conditions (NOC)

are defined so that ambient temperature is 20 OC, wind speed is 1m/s, and insolation is

800 W/m2 [4.3, 4.4].

The power produced by the panel is calculated as the product of voltage and current.

That means this power changes as a function of voltage and hence a clear maximum

value can be produced at certain output voltage. This maximum is defined as the MPP

161
of the panel. The voltage and current of the panel at this point are usually given in the

data-sheets of solar panels. This point, however, does not stay constant. This is

attributed to the fact that the voltage and current change as a function of both

insolation and temperature. As described above, insolation mostly affects current and

temperature mostly affects voltage.

Since the MPP does not remain constant in the changing operating conditions of a

solar panel. Therefore, the power converter used to extract power from the PV-module

should have some means of control to track the required MPP. This control is referred

to as maximum power point tracking (MPPT). A directly connected fixed load

resistance cannot extract maximum power from the PV system under varying radiation

and temperature conditions. To get maximum power output at all time, the PV array

should track the varying MPP. Therefore, a controller capable of tracking the

maximum power point must be implemented to regulate the interfaced converter.

There are a number of Maximum Power Point Tracking (MPPT) techniques. All of

these methods require dedicated algorithm to specify the location of the operating

point with respect to the maximum power point. In some cases, these algorithms can

deliver sub-optimum power output depending on the step size of updating panel

voltage. A good MPPT technique should accurately track the maximum power to

reduce the cost/kWh generated by the panel.

4.4 Power Conditioning Devices

The tracking of maximum power can be achieved by introducing a DC/DC converter

to adjust the panel voltage properly. The control section consists of two control loops:

one for the input and the other for the output. The PV generator voltage is regulated at

162
the required level by suitable adjustment of the DC/DC converter. In one of studied

algorithm the voltage level can be kept constant (CV algorithm) or can track the actual

MPP by an appropriate searching techniques (MPPT) [4.5].

The practical way to connect PV-array to the AC-grid is to connect some panels in

series and use a DC-boost converter to boost the voltage for the inverter as shown in

Fig. 4.5. By this way the boost converter ensures that the panel system is at its MPP

and that the inverter is fed with the appropriate voltage level of the grid. Still, enough

panels should be connected in series to improve the efficiency of the boost conversion

as high as possible [4.8].

This dual device configuration has some advantages over single stage inverter. Mainly,

a high enough DC-voltage can be produced for the inverter to function at the MPP of

the installation with high efficiency and good power quality. This can be done without

exceeding the voltage limits at the terminals of the panel installation in case of single

stage inverter [4.8-4.10].

Fig. 4.5 A PV-array with central boost converter and inverter

163
4.5 Power Electronics Devices for PV arrays

From the discussion in the previous subsection, it can be observed that the most

important power electronics device for the PV application is the DC-DC converter and

the DC-AC inverter as shown in Fig. 4.6. In general, the MPPT and voltage boost are

done by the DC-DC converter controller. The power flow control to the utility and the

sinusoidal unity power factor current-injection is achieved by the DC-AC inverter

controller [4.11-4.13].

The use of boost converters also allows for greater modularity: several boost

converters can be used to feed one inverter. If MPP-tracking is implemented in the

boost converters rather than the inverter, fewer panels will be tracked by each

converter leading to more precise tracking. A diagram of an installation with string

boost converters is presented in Fig. 4.6.

Partial shading or dirt problems can have a significant effect on energy harvesting in

large installations. If the boost converters are positioned close to the panels,

transmission lines can also be made with high voltages, which reduce losses from

cable

164
 Fig. 4.6 PV-array with string boost converters
resistance. On the other side, the DC-boost converters will introduce additional losses

by an added converter stage. Previous measurements [4.9] showed that the efficiencies

of the converter reached a value of 97% for a 100kW boost converter.

However, market leaders in solar inverter technology are claiming 98% efficiencies

for whole inverter systems [4.10].

4.6 Boost Converter

As previously stated, the maximum power point tracking is basically a load matching

problem. In order to change the input resistance of the panel to match the load

resistance (by varying the duty cycle), a DC to DC converter is required. The

efficiency of the DC to DC converter is maximum for a buck converter, then for a

buck-boost converter and minimum for a boost converter. However, the grid

connected PV array and pumping system require higher voltage that produced by the

array. Therefore, the boost converter is used to obtain the suitable voltage level as

shown in Fig. 4.7.

165
 Fig. 4.7 Circuit diagram of a Boost Converter
4.6.1 Mode 1 of the Boost Converter Operation

When the switch is closed the inductor gets charged through the input DC voltage and

stores the energy. In this mode inductor current rises (exponentially) but for simplicity

we assume that the charging and the discharging of the inductor are linear. The diode

blocks the current flowing and so the load current remains constant which is being

supplied due to the discharging of the capacitor.

Fig. 4.8 Mode 1 operation of Boost Converter

4.6.2 Mode 2 of the Boost Converter Operation

In mode 2 the switch is open and so the diode becomes short circuited as shown in Fig.

4.9. The energy stored in the inductor gets discharged through opposite polarities

which charge the capacitor. The load current remains constant throughout the

operation. The equivalent load resistance seen by the PV-panel is illustrated in Fig.

4.10.

166
 V Lavg D Vin (1 D) (Vin Vout ) 0 (1.4)

Vout V in (1.5)
1D

Fig. 4.9 Mode 2 operation of Boost Converter

Fig. 4.10 The relation between the equivalent load resistance and the duty cycle

167
To extract maximum power (130W), connect a boost converter between the panel and

the load resistor. By using D the equivalent load resistance is modified so that its seen

value from the panel corresponds to the maximum power.

Fig. 4.11 Tracking the maximum power by changing the duty cycle

The selection of the switching devices depends mainly on the operating voltage and

current of the semiconductors to be used in the power converter, as shown in Fig. 4.12.

For low power and low voltage applications, MOSFET represents the most

appropriate switching devices. The MOSFET has very low conducting losses which

results in high switching efficiency. Additionally, it has a parasitic anti-parallel diode

168
that can be used as a freewheel-diode. With higher voltage and current rating

MOSFET would be replaced by IGBT and then GTO-SCR [4.11, 4.12].

Fig. 4.12 Voltage and current rating of switching devices

Other important design parameters are inductance and capacitance. The main purpose

of the capacitance in the DC-link is to keep the input voltage to the inverter at constant

value. Its size is determined so that the voltage ripples are restricted within the

prespecified limits. The capacitor must also be able to carry a fairly high

current, roughly 1.5-2 times the mean DC current [4.13, 4.14].

The selection of the series inductance depends mainly on the required power quality

and efficiency. The magnetic material together and the value of the flux density

determine the iron losses. If the switching frequency is increased, the ripple

losses will normally decrease. On the other side the switching losses in the

transistors will increase. In fact a high inductance will reduce the ripple current and,

consequently decrease the ripple losses but will, on the other hand, increase copper

losses and, consequently, increase the weight and cost of the inductance.

169
Electrolytic capacitors will be avoided in this project due to their relatively low

lifetime especially at high temperature operation. The integrated converters must,

ideally, operate for the same lifetime as the PV modules. Therefore, ceramic

components will be used. Capacitors were chosen based on the required capacitance

value, voltage and

RMS current ratings, and temperature derating of the dielectric [4.13, 4.14]

4.7 DC-DC Converter Models

The DC-DC converter model mainly attempts to correlate the input and output loads.

The output load defines the downstream loading seen by the converter. On the other

hand, the input load represents the upstream loading that the converter places on the

PV arrays. The output voltage is kept constant as possible. Other dominant variables

include output current and voltage range.

For larger grid connected PV-arrays with a capacity over 10kW, three-phase inverters

are required. All the configurations described for single-phase utility connection can

be used for the three phase. Again, the isolation from the grid can be done either by

using a line-frequency or a high-frequency transformer. In such case, an additional

converter is required for converting the produced DC voltage from PV array into the

AC frequency of the grid.

The DC output of the array is connected across a DC-Link capacitor to a voltage-

source three-phase inverter. The output of each phase of the converters is connected to

an inductor and capacitor to limit the high-frequency harmonics injected into the grid.

170
A synthesized AC output voltage is produced by appropriately controlling the inverter

switches. A three-phase transformer is then used to connect the power to the utility.

4.8 Maximum Power Point Tracking Algorithms

The general block diagram of Maximum power Point Tracking algorithms for the PV

array is given in Fig. 4.13. The basic concept of MPPT depends mainly on the applied

control technique to move the operating point toward the maximum power [4.15-

4.19].

Fig. 4.13 Concept of maximum power point tracking

The MPPT concepts can be classified into two groups as follows:

Indirect MPP trackers:

This type of MPP tracker estimates the MPP voltage by means of simple measurements.

171
 1. The operating voltage of the solar generator can be adjusted seasonally. Higher

MPP voltages can be expected in winter because of lower cell temperatures.

2. The operating voltage can be adjusted according to the module temperature.

3. The operating voltage can be derived from the instantaneous open-circuit voltage

by multiplication with a constant factor, for example, 0.8 for crystalline silicon

solar

cells.
These indirect MPP trackers results in an approximate but in many cases acceptable

optimum operation point and defined as Constant Voltage algorithms.

Direct MPP trackers

In these trackers, the optimum operating voltage is determined from measured

currents, voltages or the actual power produced by the PV array. Therefore, they are

able to respond to quick environmental changes. These algorithms are defined as

Incremental Conductance trackers. In other direct trackers, the operating voltage is

periodically changed in small steps until the maximum power is approached. The

increment can either be constant or can be updated when searching the optimal

operating point. If the module’s power increases from one step to the next, the search

direction is retained: otherwise it is reversed. In this way, the MPP is found and the

operating point oscillates around the actual MPP. These trackers are defined as Perturb

and Observe algorithms [4.17-4.20]

Table 4.1 summarized the above mentioned MPP tracking algorithms studied in this

report according to the local environmental conditions in Kuwait.

172
 Table 4.1 Maximum power point tracking algorithms
Indirect Direct

Constant Voltage Incremental Conductance

Perturb and Observe

4.8.1 Constant Voltage Algorithm

This simple and robust method depends mainly on the Open Circuit Voltage of the

relay as dominant parameter for tracking the maximum power point. The ratio

Vmpp/Voc is usually kept constant by using this algorithm [4.21]. This method was

proposed because the maximum power point Voltage does not change so much with

radiation variations. The operating point of the PV array is kept near the MPP by

regulating the array voltage and matching it to a fixed reference voltage V ref. The Vref

value is set equal to the V MPP of the PV module. This method assumes that individual

insulation and temperature variations on the array are insignificant, and that the

constant reference voltage is an adequate approximation of the true MPP. This means

that CV algorithm is based on approximately constant ratio between V MPP and Voc as

given in (4.6):

VMPP ' kVoc (1.6)

By the adjusting of array voltage to this calculated value, the operation at MPP is

achieved. In literature, K value ranges between 73% and 80% [4.22-4.25]. The

measurement of the voltage VPV is necessary in order to set up the duty-cycle of the

173
DC/DC converter. After determining the value for V MPP, the DC/DC boost converter

is switched to quickly force the array terminal voltage to its MPP value. The

difference between MPP fraction voltage and the array output represents the error of

the PID controller in Fig. 4.14. This PID controller regulates the PV array voltage by

continuously switching on and off of the IGBTs at high frequency according to the

varying duty cycle. The fast switching action of the boost converter decouples the

dynamics of the PV array due its changes in voltage or current from that of the DC

link capacitor, which offers good performance under changing weather conditions.

The other

Fig. 4.14 Boost with MPP block diagrams

advantage of using the boost converter is to increase the array voltage to higher levels

suitable for grid interconnection.

4.8.2 Perturbation and Observation (P&O) Method

The P&O method has been widely used for MPPT because it is straight forward. This

approach requires only measurements of Vpv and Ipv, and it can track the MPPT

accurately under varying irradiance and temperature. According to its name, the P&O

algorithm perturbs Vpv and observes the impact of this change on the output power of

174
the PV array [4.22, 4.26]. Fig. 4.15 shows the flow chart of the P&O algorithm. At

each time interval (k), the power of the array Ppv(k) as a function of Vpv and Ipv is

calculated. The calculated value of Ppv(k) is compared to the value Ppv(k-1)

determined at the previous interval. If the output power has found to be increased, Vpv

is adjusted further in the same direction as in the previous interval. Otherwise if the

output power has decreased, then Vpv should be perturbed in the opposite direction of

the previous interval. According to this procedure the MPP is thus perturbed at cycle.

The flow chart of algorithm has 4 cases as shown in Fig. 4.15 and can be detailed as

following

( When ΔP <0 & V(j)>V(j-1), this yields to Vref =V (j+1) = V (j) - ΔV

( When ΔP <0 & V(j)<V(j-1), this yields to Vref =V (j+1) = V (j) + ΔV

( When ΔP >0 & V(j)<V(j-1), this yields to Vref =V (j+1) = V (j) - ΔV

( When ΔP >0 & V(j)<V(j-1), this yields to Vref =V (j+1) = V (j) + ΔV

When the MPP is approached, Vpv may oscillate around the optimal value Vpv,mpp

depending on the magnitude of the step size. These oscillations lead to a reduction in

the produced power of the relay with increasing the step size of the perturbation. In

this respect it should be noted, if this step size is large, the MPPT algorithm responds

quickly to sudden changes in environmental conditions. On the other side, if the step

size is small the algorithm becomes relatively slow and not be able to respond quickly

to rapid changes in temperature or irradiance.

Fig. 4.16 shows the block diagram of the automatic converter control to reach the

MPP. This control includes a PI controller. The error signal driving the PI controller is

175
equal to the differences between the adjusted reference terminal voltage, Vref and the

array voltage Vpv. The output of the controller is fed to PWM to generate the firing

signals of the DC/DC converter.

4.8.3 Incremental Conductance Method

The Incremental Conductance method has been proposed to avoid the drawbacks of

the P&O MPPT method. It is based on the fact that the derivative of the output power

Ppv with respect to the panel voltage Vpv is equal to zero at the maximum power point

[4.26-4.28]. The PV panel Ppv-Vpv characteristic in Fig. 2.2 shows further that this

176
177
 Fig. 4.15 Block diagram of P&O methods

Fig. 4.16 Schematic diagram of Converter control for MPPT

derivative is positive to the left of the maximum power point and negative to the right
of maximum power point. This leads to the following set of equations:
dPpv (1.7)
dIV ( pv pv ) I pv V pv dI pv
0
at MPP dV pv dV pv
dV pv
dPpv
dIV ( pv pv ) I pv V pv dI pv 0 to the (1.8)
dV pv left of MPP dV pv dV pv
dPpv
dIV ( pv pv ) I pv V pv dI pv
0 to the
dV pv right of MPP dV pv dV pv (1.9)
These equations can be written as:

dI pv I pv at MPP (1.10) dV pv V pv

pv pv
to the left of MPP (1.11)
dV pv V pv
dI I
dI pv
I pv to the right of MPP (1.12)
dV pv V pv

178
The equation above can be used as a control algorithm to control the inverter operating

point by measuring the incremental and instantaneous inverter

dV V
dIPV IPV
conductance ( and , respectively). The schematic

diagram of the control

PVPV

algorithm is shown in Fig. 4.17. The method may be more difficult to implement than the

previous methods because it involves cases in which the denominators may be equal to

zero.

179
 Fig.

4.17 Schematic diagram of incremental conductance MPPT algorithm Vref is the

reference voltage at which the PV array is forced to operate. For MPP, the error
signal (e) determined in Fig. 4.18 is fed to PI controller to regulate the array terminal
voltage to the MPP value by switching the IGBT devices according to the duty cycle
of the converter. Once the MPP is reached, the operation of the PV array is
maintained at this point unless a change in dI is noted, indicating a change in
atmospheric conditions and the MPP. The algorithm decrements or increments Vref
to track the new MPP.
By proper control of the power converter, the initial operating point is set to match a load

resistance proportional to the ratio of the open-circuit voltage (VOC) to the shortcircuit

current (ISC) of the PV array. A less obvious, but effective way of performing the IC

180
technique is to use the instantaneous conductance and the incremental conductance to

generate an error signal (e) as shown in Fig. 4.18.

I dI
e
(1.13)
V dV

Fig. 4.18 Simulink model to determine the error signal in IC algorithm

4.8.4 Comparison of MPPT Algorithms

Simulink software is used to model the studied grid-connected PV array with a MPPT

boost converter tracker as shown in Fig. 4.19. The solar array block simulates the

nonlinear V–I characteristics of the panel at different solar radiations and temperatures.

The specifications of studied solar panel are given in Table 4.2. As depicted in Fig. 4.19

the boost converter is connected from the output of the solar array to the DC/AC inverter.

The produced electric power is controlled by varying the on/off duty cycle of the

converter.

181
 Fig. 4.19 Simulated Grid Connected Solar Array

Table 4.2 PV Panel Specifications

Description Symbol Value Unit

Open Circuit Voltage VOC 20.8 V

Short Circuit Current ISC 3.6 A
O.C. Voltage/Temp. Coff. KV -75x10-3 V/oC
No. of Series cell/module NS 36
No. of series connected module NSS 33
No. parallel module NP 60
Array power rating at NOTC PA 100 kW

In order to verify the MPP tracker for the photovoltaic simulation system, the described

MPPT techniques are compared at different ambient conditions in Kuwait to show how

they can effectively and accurately tracks the maximum power of the array. The simulation

is carried using MATLAB/SIMULINK. The output of the MPPT control block is the

gating signal which is used to drive the IGBT switching devices.

The boost converter is designed according to the maximal power of the array. Table 4.3

shows the parameters of the DC-DC boost converter. The test data consist of hourly

radiation and temperature for time horizon of 6 years in Kuwait as described in the

previous report. This data are applied to check the effectiveness of the described MPPT

techniques in producing the maximum annual energy of the photovoltaic systems.

182
 Table 4.3 Buck-Boost Converter Parameters
Parameter Value

Inductance L 5 mH

Capacitance C1 100 μF

Capacitance C2 6000 μF

Frequency 5 KHz

4.9 Simulation and Numerical Results

For each MPPT technique and for hourly input data, the energy supplied by the PV system

was calculated for each month averaged over the studied period of 6 years. The

implemented controller for the three studied MPP algorithms is PI type with K p=2 and

Ki=0.05. The same Boost converter with its maximum MPP controller is used for

comparison. The results are summarized in Table 4.4. The simulation results indicated that

the generated energy from the tracking algorithm using P&O MPPT is maximum with

annual energy of 166233.548 kWh. From the result in Table 4.4, It can be noted that the

P&O and IC algorithms over the year are superior to the CV algorithm. The produced

energy using CV technique is the lowest of the three studied MPPT methods with annual

value of 165012.804. This is attributed to the fact, that CV technique does not follow the

MPP, but instead fixes the reference voltage to the optimal voltage under STC or to

another best fixed voltage, holding it constant under any operating condition. However, in

December and January the generated energy using the CV technique is slightly greater

than both IC and P&O techniques. On the other hand, IC has slightly smaller energy

production compared to P&O techniques in February, March, April, May August,

183
September, October and November. On the other hand, the generated energy in June and

July using IC MPP technique is maximum compared to that of CV and P&O techniques.

Table 4.4 Energy Generated Using MPP Techniques in kWh

Month CV P&O IC
Jan 6402.368 6397.408 6395.114
Feb 7492.296 7521.864 7440.720
Mar 13050.194 13031.284 13019.318
Apr 14500.200 14522.400 14477.880
May 19403.458 19533.100 19489.328
Jun 20588.460 20765.760 20779.740
Jul 21131.274 21385.784 21399.920
Aug 19901.690 20234.444 20183.914
Sep 16430.160 16626.240 16577.160
Oct 12045.360 12153.302 12120.256
Nov 7418.340 7432.860 7430.340
Dec 6649.004 6629.102 6618.562
Total 165012.804 166233.548 165932.252

Moreover, for these techniques the choice of sampling period is very critical; if the period

is too short, energy production will be very low because of the increased number of

electronic switching. If the period is too long, on the other hand, the MPP cannot be

closely followed when rapid irradiance variation occurs.

The purposes of the next tests are to investigate the dynamic characteristics of a PV system

and to calculate the amount of power, voltage and current using the described MPPT

controller. It should be noted that the generated power of has the same shape as the solar

insulation input, the only difference is a small transient from the rapid insulation variation

by using P&O and IC techniques. Comparing the output array voltage, it can be observed

that the CV method is more stable with solar insulation variation.

184
In particular due to lack of space, Figs. (4.20-4.22) show the selected responses of the PV

array using the IC, P&O and CV algorithms. In the first test, the MPPT controller is tested

under a sunny day, and then cloudy and partially cloudy days to calculate the amount of

the oscillation in PV operating points. The dynamic response of the PV array indicated that

the CV technique delivers the more stable voltage, current and power waveforms with

negligible oscillation amplitudes. Table 4.5 summarizes the oscillations in voltage, current

and power using IC and P&O techniques. This is attributed to the IGBTs switching effect

to change the array voltage to satisfy the MPP conditions of IC and P&O techniques

4.9.1 Dynamic response of the PV array for clear day 28th June

)
(a Radiation (b) Temperature
Fig. 4.20 Radiation & Tempreture (26th June)

(a) Power (b) Current (c) Volt

185
 Fig. 4.21 Dynamic Response of PV Array using CV algorithm

)
(a) Power (b) Current (c Volt
Fig. 4.22 Dynamic Response of PV Array using IC algorithm

(a) Power (b) Current (c) Voltage

Fig. 4.23 Dynamic Response of PV Array using P&O algorithm

4.9.2 Dynamic response of the PV array for partially clouded day 19th

Jan

186
 (a) Radiation (b) Temperature
Fig. 4.24 Radiation & Tempreture (19th January)

(a) Power (b) Current (c) Volt

Fig. 4.25 Dynamic Response of PV Array using CV algorithm

)
(a) Power (b) Current (c Volt
Fig. 4.26 Dynamic Response of PV Array using IC algorithm

(c) Voltage
(a) Power (b) Current
Fig. 4.27 Dynamic Response of PV Array using P&O algorithm

187
4.9.3 Dynamic response of PV array for clouded day 16th Nov

(a) Radiation (b) Temperature

Fig. 4.28 Radiation & Tempreture (16th Novemper)

(a) Power (b) Current (c) Volt

Fig. 4.29 Dynamic Response of PV Array using CV algorithm

(a) Power (b) Current (c) Volt

Fig. 4.30 Dynamic Response of PV Array using IC algorithm

188
 (a) Power (b) Current (c) Voltage
Fig. 4.31 Dynamic Response of PV Array using P&O algorithm

Table 4.5 PV Performance under different sky conditions

Sky Unit P&O IC

Condition P-P Oscillation Max Value P-P Oscillation Max Value

28th June Voltage (V) 26 510 59 510

Clear Current (A) 13 272 15 273

Power (Kw) 0.04 140 0.03 141

19th Jan Voltage (V) 31 481 79 480

Partially Current (A) 2 59 1.5 58.5

Clouded Power (Kw) 0.05 28.5 0.08 28

16th Nov Voltage (V) 36 479 53 477

Clouded Current (A) 1.7 42.4 2.5 42

Power (Kw) 0.07 22.3 0.1 22.2

189
4.10 Conclusion

This chapter presents in details a comparative study between constant voltage technique

and the two most popular algorithms which are Perturb & Observe algorithm and

Incremental Conductance algorithm. To carry out this study the hourly solar radiation, and

temperature were collected for a period of 6 years in Kuwait. The Boost converter with the

associated PI controller is used for this comparison. Matlab/ simulink have been used for

modeling and simulation of the PV array. The capacity of the simulated PV array is 100

kW under 1000W/m2 and 250C conditions. A ranking was proposed of the different MPPT

techniques based on the sum of the energy generated under different weather conditions.

Due to lack of space the average generated energy each month was displayed and

compared.

The results indicated that the P&O and IC algorithms are in general more efficient than CV

MPPT technique. In fact, the CV technique did not completely follow the MPP, but

instead fixed the reference voltage to the optimal voltage under STC values, holding it

constant under any operating condition. The P&O technique provides the greatest annual

energy supply for the studied period of 6 years. This is confirmed by their widespread use

in commercial implementations. The IC method provided only slightly less annual energy

than the P&O. The generated power output of the three methods has the same shape as the

solar insulation. The only difference was a small transient from the rapid insulation

variation.

In addition, the dynamic response comparison of voltage, current and power output has

been traced under different condition of available radiation and temperature. The MPPT

controller was tested under a sunny day, and then cloudy and partially cloudy days to

190
 calculate the amount of the oscillation in PV operating points. The simulation results

indicated that the CV techniques delivered more stable voltage and current signal of the

PV array compared to IC and P&O techniques. Underway the fuzzy logic controller will

be implemented to enhance the performance of the most suitable P&O algorithm.

References

[4.1] T. Nordmann and L. Clavadetscher, “Understanding tempreture effect on pv system

performance”, Proceedings of 3rd World Conference on Photovoltaic Energy
Conversions, vol. 3, May 2003, pp. 2243-2246.
[4.2] Photovoltaic devices – part 3: Measurement principles for terrestrial Photovoltaic
(PV) solar devices with reference spectral irradiance data, IEC 60904-3-ED. 2.0,
2008.
[4.3] BP, 230W Photovoltaic module BP 3230T. [Online]. Available:
http://www.bp.com/liveassets/bp_internet/solar/bs_solar _usa/STAGING/local_
assets/downloads_pdfs/pq/BP3230T_1-10.pdf
[4.4] T. Markvart (edit.), Solar Electricity (Energy Engineering Learning Package).
Chichester, England: John Wiley & Sons, 1994.
[4.5] F. A. Farret and M. G. Simoes, Integration of Alternative Sources of Energy. Hoboken,
NJ: Wiley-IEEE press, 2006.
[4.6] T. Esram and P. Chapman, “Comparison of Photovoltaic array maximum power point
tracking techniques,” IEEE Transactions on Energy Conversion, Vol. 22, no. 2, Jun.
2007, pp. 439-449.
[4.7] T. L. Nguyen and K.-S. Low, “A global maximum power point tracking scheme
employing direct search algorithm for Photovoltaic systems,” IEEE transactions on
Industrial Electronics, vol. 57, no. 10, Oct. 2010, pp. 3456-3467.
[4.8] Y. Huang, F. Peng, J. Wang, and D. Yoo, “Survey of the power conditioning system
for PV power generation,” in Power Electronics Specialists Conference, 2006.
PESC’06. 37th IEEE. IEEE, 2006, pp.1-6.
[4.9] M. Liukkonen, A. Hentunen, and J. Suomela, \Analysis of the ultracapacitor module
in power buffering," in Proceedings of 4th European Symposium on Super
Capacitors & Applications, Bordeaux, France, Oct. 2010.
[4.10] SMA, Sunny Mimi Central Installation guide. [Online]. Available:
http://download.sma.de/smaprosa/dateien/5710/SMC9-11TLRP-IEN100640.pdf
[4.11] Texas Instruments. Understanding Boost Power Stage in Switchmode Power Supplies.
Texas Instruments, Mar. 1999. PDF
[4.12] Van Duijsen P.J., Bauer P., Lascu D., "Selection of Semiconductor Models in Power
Electronics Simulation", Proceedings of the PCIM 2003

191
[4.13] R. W. Erickson and D. Maksimovic. Fundamentals of Power Electronics. Kluwer
Academic Publishers, 2nd edition, 2001.
[4.14] "Boost Switching Converter Design Equations." Electronics, PCB Design and PCB
Layout. Day counter, Inc, 2004. Web. 25 Feb. 2011.
<http://www.daycounter.com/LabBook/BoostConverter/Boost-
ConverterEquations.phtml>.
[4.15] C.Hua, C.Shen, " Comparative study of peak power tracking techniques for solar
storage system” Applied Power Electronics Conference and Exposition, APEC 98.
Conference Proceedings 1998,. Volume 2, pp 679-685, 15-19 Feb. 1998.
[4.16] Trishan Esram , Pa trick L. Chapman, “Comparison of photovoltaic array maximum
power point tr acking techniques”, IEEE Transactions on energy conversion, Vol.
22, No. 2, June 2007.
[4.17] N. Femia, G. Petrone, G. Spagnuolo, and M. Vitelli, “Optimization of perturb and
observe maximum power point trackingmethod,” IEEE Trans. Power Electron., vol.
20, no. 4, pp. 963–973, Jul. 2005.
[4.18] Weidong Xiao, William G. Dunford, Patrick R. Palmer, Antoine Capel, “Regulation
of photovoltaic voltage”, IEEE Transactions on Industrial Electronics, Vol. 54, No.3,
June 2007
[4.19] Trishan Esram , Pa trick L. Chapman, “Comparison of photovoltaic array maximum
power point tr acking techniques”, IEEE Transactions on energy conversion, Vol.
22, No. 2, June 2007
[4.20] Calavia, M., Perié, J.M., Sanz, J.F. and Sallán, J., 2010. Comparison of MPPT
strategies for solar modules. International Conference on Renewable Energies and
Power Quality, Granada, Spain, March 23-25, 2010.
[4.21] Y. M. Tung, A. P. Hu, N. K. Nair, “Evaluation of Micro Controller Based Maximum
Power Point Tracking Methods Using dSPACE Platform,” Australian University,
Power Engineering Conference 2006.
[4.22] F. Liu, Y. Kang, Y.Zhang and S. Duan, “Comparison of P&O and Hill Climbing
MPPT Methods for Grid-Connected PV Converter,” Industrial Electronics and
Applications, 2008. ICIEA 2008. 3rd IEEE Conference, pp. 804 - 807.
[4.23] S. Jain, and V. Agarwal, “Comparison of the performance of maximum power point
tracking schemes applied to single-stage grid-connected photovoltaic systems,” IET
Electr. Power Appl., 2007, pp. 753 - 762.
[4.24] C. Jaen, C. Moyano, X. Santacruz, J. Pou, and A. Arias, “Overview of maximum
power point tracking control techniques used in photovoltaic systems,” Electronics,
Circuits and Systems, 2008. ICECS 2008. 15 th IEEE International Conference, pp.
1099 - 1102.
[4.25] D. P. Hohmand , M. E. Ropp, “Comparative Study of Maximum Power Point
Tracking Algorithms,” Progress in Photovoltaic Research and Applications, vol.
11, no. 1, pp. 47–62, 2003
[4.26] Y.-C. Kuo, T.-J. Liang, and J.-F. Chen, "Novel maximum-power-point tracking
controller or photovoltaic energy conversion system," IEEE Trans. Ind. Electron.,

192
 vol. 48, pp. 594-601, June 2001.
[4.27] G. J. Yu, Y. S. Jung, J. Y. Choi, I. Choy, J. H. Song, and G. S. Kim, "A novel two-
mode MPPT control algorithm based on comparative study of existing algorithms,"
in Conf. Record of the Twenty-Ninth IEEE Photovoltaic Specialists Conf., 2002, pp.
1531-1534.
[4.28] Hairul Nissah Zainudin, Saad Mekhilef, Comparison Study of Maximum Power
Point Tracker Techniques for PV Systems, Proceedings of the 14th International
Middle East Power Systems Conference (MEPCON’10), Cairo University, Egypt,
December 19-21, 2010, Paper ID 278..

193