1 MASTER EN SISTEMAS DE ENERGÍA ELECTRICA Trabajo Fin de Master Escuela Técnica Superior de Ingenieros (ETS) Departamento de Ingeniería Eléctrica Universidad de Sevilla Maximum Power Point Tracking in Photovoltaic System by Master Student: Ahmed Mohamed Abd el Motaleb Tutor: Dr. Antonio de la Villa Jaen
186
Embed
Maximum Power Point Tracking in Photovoltaic Systembibing.us.es/proyectos/abreproy/70172/fichero/Master+Work.pdf · Maximum Power Point Tracking in Photovoltaic System This thesis
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
MASTER EN SISTEMAS DE ENERGÍA ELECTRICA
Trabajo Fin de Master
Escuela Técnica Superior de Ingenieros (ETS)
Departamento de Ingeniería Eléctrica
Universidad de Sevilla
Maximum Power Point Tracking in
Photovoltaic System
by
Master Student: Ahmed Mohamed Abd el Motaleb
Tutor: Dr. Antonio de la Villa Jaen
2
ABSTRACT
Maximum Power Point Tracking in Photovoltaic System
This thesis provides theoretical studies of photovoltaic . also we concentrate
on different types of maximum power point tracking (MPPT) including its
circuit components and function of each component. the thesis includes
discussion of various MPPT algorithms and control methods. Most popular
MPPT methods had been explained through this thesis in simplicity and
details with different ideas and mentalities .
Also through the last chapter the design and simulation of a simple but efficient photovoltaic system had been introduced including tests and
comparisons of different MPPT systems with values of each method .
And finally we end with conclusions will explain in brief the different
between each method and MPPT system requirements .
3
Acknowledgment
I would like to first acknowledge my calm and respectable professor,
Antonio de la Villa, for his support and advice throughout this thesis. His
guidance and dedication gave me good experience during the course.
I would also like to express my sincere appreciation to rest of my master
professors, for their valuable informations and feedback through last year
I would like to thank my friends Mohamed Abd El.Twab , Amr Ismail for
their support and ideas generated from our numerous discussions to be
incorporated through this thesis.
Finally, to my parents, my sister, and my friends .
This chapter concentrates on benefits of photovoltaic system and its
superiority over other conventional power sources , applications of
photovoltaic system .
Also we had explained here the MPPT problem and its necessity in PV system
And end by components of MPPT control system
13
1.1) INTRODUCTION
All electricity generation technologies generate carbon dioxide (CO2) and other greenhouse gas emissions. To compare the impacts of these different technologies accurately, the total CO2 amounts emitted throughout a system’s life must be calculated. Emissions can be both direct – arising during operation of the power plant, and indirect – arising during other non-operational phases of the life cycle. Fossil fuelled technologies (coal, oil, gas) have the largest carbon footprints, because they burn these fuels during operation. Non-fossil fuel based technologies such as wind, photovoltaics (solar), hydro,biomass, wave/tidal and nuclear are often referred to as ‘low carbon’ or ‘carbon neutral’ because they do not emit CO2 during their operation. However, they are not ‘carbon free’ forms of generation since CO2 emissions do arise in other phases of their life cycle such as during extraction,
construction, maintenance and decommissioning .
14
What is a carbon footprint? A ‘carbon footprint’ is the total amount of CO2 and other greenhouse gases, emitted over the full life cycle of a process or product. It is expressed as grams of CO2 equivalent per kilowatt hour of generation (gCO2eq/kWh), which accounts for the different global warming effects of other greenhouse gases.
Life cycle CO2 emissions for electricity generation technologies
15
Range of carbon footprints for UK & European ‘low carbon’ technologies
PHOTOVOLTAIC HISTORY
The history of PV dates back to 1839 when a French physicist, Edmund
Becquerel, discovered the first photovoltaic effect when he illuminated a
metal electrode in an electrolytic solution . Thirty-seven years later British
physicist, William Adams, with his student, Richard Day, discovered a
photovoltaic material, selenium, and made solid cells with 1~2% efficiency
which were soon widely adopted in the exposure meters of camera [16].
In 1954 the first generation of semiconductor silicon-based PV cells was
born, with efficiency of 6% , and adopted in space applications. Today, the
16
production of PV cells is following an exponential growth curve since
technological advancement of late ‘80s that has started to rapidly improve
efficiency and reduce cost.
Recent awareness of global warming and increasing prizes of fossil fuels have
drawn more attention towards the usage of renewable energy sources today.
among the various renewable energy systems, solar energy systems have the
merits such as clean without any environmental pollution problems and
infinite in mass, and are becoming one of the future energies.
17
PV powered, Diesel powered, vs. Windmill
1.2) APPLICATION OF PHOTOVOLTAIC
Following figures show different application of PV
18
Hybrid Power System
19
Desalination Plants
20
Heating & Residential Use
21
Communication Systems
22
Street Lighting
23
1.3) MPPT PROBLEM
The amount of power generated from a photovoltaic (PV) system mainly
depends on the following factors, such as temperatures and solar irradiances.
according to the high cost and low efficiency of a PV system, it should be
operated at the maximum power point (MPP) which changes with solar
irradiances or load variations. number of maximum power point tracking
(MPPT) techniques have been developed for PV systems .
24
and the main problem is how to obtain optimal operating points (voltage and
current) automatically at maximum PV output power under variable
atmospheric conditions.
The majority of MPPT control strategies depend on characteristics of PV panels
in real time, such as the duty cycle ratio control and using a look-up table.
25
26
27
1.4) MPPT STRUCTURE
Structure of PV Power System
28
Current Sensor
Voltage Sensor
29
DC-DC CONVERTER
DC-DC Converter in MPPT system is either
1) Buck Converter
2) Boost Converter
3) Buck-Boost Converter
All will be depending on the batteries we will store energy in .
1) Buck Converter
30
Buck Converter produces voltage equal or lower than the input voltage
D : the duty cycle ratio of converter
2) Boost converter
Boost converter produces output voltage that is greater or equal to the input
voltage.
31
3) Buck-Boost Converter
Output of a buck-boost converter either be higher or lower than
the source voltage.
–If D>0.5, output is higher
–If D<0.5, output is lower
32
DC-AC Inverter
The transistor switching signals for the inverter are obtained from the real and
the reactive power control system .
33
The controller controls the phase angle and amplitude of the voltage across
the transformer The difference in the phase angle between the voltage across
the transformer and the utility side voltage determines the direction of the
real and reactive power flow .
P-Q CONTROL
The P-Q controller basically consists of PI controllers to control the phase angle
and the modulation index. The main requirement for the inverter switching
signals is the phase angle and amplitude of the inverter voltage.
The real power is directly proportional to the phase angle if the angle is small.
Hence the real power flow can be used to control the phase angle of the
inverter , while reactive power flow is controlled by the amplitude of
transformer voltage .
34
35
MPPT Controller
Analog controllers have traditionally performed control of MPPT.
however, the use of digital controllers is rapidly increasing because they
offer several advantages over analog controllers. First, digital controllers
are programmable thus capable of implementing advanced algorithm with
relative ease. It is far easier to code the equation, x = y × z, than to
design an analog circuit to do the same . For the same reason, modification
of the design is much easier with digital controllers. They are immune to
time and temperature drifts because they work in discrete, outside the linear operation. As a result, they offer long-term stability. They are also
insensitive to component tolerances since they implement algorithm
in software, where gains and parameters are consistent and reproducible .
they allow reduction of parts count since they can handle various tasks in a
single chip. Many of them are also equipped with multiple A/D converters
and PWM generators, thus they can control multiple devices with a single
controller.
36
37
CHAPTER “ 2”
38
Introduction to Chapter 2
Through this chapter various and different methods of MPPT which are
considered the most popular MPPT methods belong to other investigators
had been explained in details and simplicity and by end of each method we
make conclusion to well understand the difference between them .
39
2.1) PERTURBATION&OBSERVATION [1],[2],[3],[4],[5]
Incrementing (decrementing) the voltage increases (decreases) the power
when operating on the left of the MPP and decreases (increases) the power
when on the right of the MPP. therefore, if there is an increase in power, the
subsequent perturbation should be kept the same to reach the MPP and if
there is a decrease in power, the perturbation should be reversed. This
algorithm is summarized in TABLE I the process is repeated periodically until
the MPP is reached. the system then oscillates about the MPP.
40
The oscillation can be minimized by reducing the perturbation step size.
however, a smaller perturbation size slows down the MPPT. A solution
to this conflicting situation is to have a variable perturbation size that gets
smaller towards the MPP
Characteristics of PV Power Curve
41
42
EXPERIMENTAL SET UP
1) DC-DC CONVERTER :
In order to obtain comparable results, it has been realized a single
device constituted by a dc-dc converter and other components able to
implement all the different MPPT techniques here analyzed, including
Open Circuit Voltage (OV) and Short Current Pulse (SC) which required
to insert further static switches to open the circuit or to create the
short-circuit condition. All the MPPT techniques here described are
easily obtained changing the software compiled on the microcontroller.
43
The control board is constituted by all the components that need for the
implementation of the various MPPT algorithms The microcontroller, in
thiscase a Microchip dsPIC30f4012, is the core of the control board.the
command connection to the power board is provided by means of driver
circuits which allow the valves commutation. the boost section is realized by
the two accumulation units, L and C out, by the T1 static switch and by the D3
diode. moreover, diode D1 is put into the circuit to protect the PV-panel against
negative current which could damage it.the measures of the PV-panel voltage,
VPV, and current,IPV, are obtained by inserting the voltage transducer V and
the current one A in the circuit as reported in following fig. showing the circuit
elements Tv0, Tsc, K1, K2, Cin and D2, that have been inserted to:
• measure the PV-panel open circuit voltage, that is necessary in OV technique,
through the opening of Tv0 valve, in this case D2 is short-circuited through K2
44
• measure the PV-panel short-circuit current, that is necessary in the SC
technique, through the closure of Tsc valve, in this case Tv0 is short-circuited
through K1.
During the tests of other MPPT techniques, the valve Tsc is kept open, while
Tv0 and D2 are short-circuited, respectively through K1 and K2 switches, to
increase converter efficiency boost
It is important to note that in the SC MPPT technique it is necessary to insert
the D2 diode to avoid, during the short-circuit test, the discharging of Cin
placed at boost input. such capacitor is always inserted in each techniques
analysed to limit the high frequency harmonic components.
The dc-dc converter is designed to work at the MPP with a duty cycle of 25%.
the dc-dc converter sizing, with a security margin, leads to the following data:
switching frequency of 20 kHz, nominal current of 15 A, and nominal voltage of
150 V. The IGBT IRG4PC30KD electronic valves are chosen . these components
45
have integrated a ultrafast recycling diode and present small switching losses
also in presence of high switching frequency.
B) PV PANEL :
The PV panels here considered are the poly crystalline 70 W
2) SOLAR SIMULATOR :
The solar simulator used in the present tests is realized by using
incandescent and halogen lamps. The maximum power of the solar
simulator is 2.8 kW and its size is 1200 mm long and 600 mm wide.
Combining the lamps, it is possible to have four different irradiation
levels equal to 0 W/m2, 272 W/m2, 441 W/m2 and 587 W/m2.
46
VERY IMPORTANT NOTE :
The previous experimental set up is valid for the following MPPT methods :
1) P & O 3) Open circuit voltage
2) Incremental Conductance 4) Short circuit current
In order to realize a precise analysis of the performance of the different MPPT
techniques, they are experimentally compared taking into account two
different irradiation diagrams. The first one, Case 1 , is characterized by
medium and medium-high irradiation levels of 441 W/m2 and 587 W/m2 with
a time of 180 s and the second one, Case2 , with low, low-medium, medium-
high irradiation levels of 0 W/m2, 272 W/m2, 441 W/m2 and 587 W/m2, with
a time of 160 s (Case 2 include a 10 s interval without irradiation).
47
48
PERTURBATION & OBSERVATION TEST :
The conventional P&O technique used in this comparison increases or
decreases the duty-cycle of =1.6% each 200 ms. It performs very well with low
radiance values: in this condition the P-V curve is very smooth near the
maximum and hence the 1.6% duty-cycle variations do not imply significant
output power reduction under steady state condition. In case of higher
irradiance values, instead, oscillations are more evident. To reduce the
oscillations it is necessary to reduce the , but this implies a reduction of the
technique’s speed during the variations. The chosen value is a compromise
between the reduction of steady state oscillations and the dynamic behaviour
of this technique.
49
SIMULATION RESULTS
50
51
2.2) P & O WITH MODIFICATION [3],[4],[5]
Also in the case of modified P&O technique, the algorithm increases or
decreases the duty-cycle with the same logic of conventioanal P&O, and
performs an iteration every 200 ms. In this technique the amplitude of duty-
cycle (increase or decrease) is proportional to the ratio dP / dV and it ranges
from 0.5% to 2.7%. The modified P&O logic with variable step is able to
reduce steady state oscillations and, at the same time, to provide higher
response speeds at medium-high irradiance level with respect to the
conventional P&O approach with fixed . This technique is very slow in
reaching MPP when irradiance level is low because dP / dV is small.
52
DISADVANTAGES OF (P & O) METHOD
Hill climbing and P&O methods can fail under rapidly changing atmospheric
conditions as illustrated in corresponding figure. Starting from an operating
point A, if atmospheric conditions stay approximately constant, a perturbation
ΔV in the PV voltage V will bring the operating point to B and the perturbation
will be reversed due to a decrease in power. However, if the irradiance
increases and shifts the power curve from P1 to P2 within one sampling period,
the operating point will move from A to C. This represents an increase in
power and the perturbation is kept the same.
the operating point diverges from the MPP and will keep diverging if the
irradiance steadily increases. To ensure that the MPP is tracked even under
sudden changes in irradiance, uses a three-point weight comparison P&O
method that compares the actual power point to two preceding ones before a
decision is made about the perturbation sign. the sampling rate is optimized .
53
2.3) INCREMENTAL CONDUCTANCE[5],[6],[7],[8]
The incremental conductance method is based on the fact that the slope of
the PV array power curve is zero at the MPP, positive on the left of the MPP,
and negative on the right, as given by :
dP/dV = 0, at MPP
dP/dV > 0, left of MPP
dP/dV < 0, right of MPP.
Characteristics of PV Power Curve
54
Since
The MPP can thus be tracked by comparing the instantaneous conductance
(I/V ) to the incremental conductance (ΔI/ΔV ) as shown in the following
flowchart in. Vref is the reference voltage at which the PV array is forced to
operate. At the MPP, Vref equals to VMPP. Once the MPP is reached, the
operation of the PV array is maintained at this point unless a change in ΔI is
noted, indicating a change in atmospheric conditions and the MPP. The
algorithm decrements or increments Vref to track the new MPP.
55
56
INCREMENTAL CONDUCTANCE TEST
The IC technique performs the test on the incremental conductance every
200 ms . This algorithm should run faster, but in this comparison is enforced
to have the same duty cycle =1.6% each 200 ms. with this variation speed
the performance is different until to arrive in steady state conditions. Case II
shows the main disadvantage of the IC technique: for low radiance values the
technique works on a P-V curve with a derivative close to zero in a large
interval around the maximum value, therefore it is not able to properly
identify the MPP. It results in oscillations around the MPP with a reduced
output energy value .
57
SIMULATION RESULTS
58
59
CONCLUSION
1) Through medium and high insolation the incremental conductance is
more effective than P & O method because incremental takes in
consideration the change in current so it will be more sensitive to any
variation in insolation.
2) Through low insolation there is no difference between P & O and
Incremental Conductance because P & O will proceed large perturbation
every sample ,and Incremental will take in consideration the slope of of
P-V curve which is actually very little at low insolations.
60
2.4) FRACTIONAL OPEN CIRCUIT VOLTAGE [5]
The near linear relationship between VMPP and VOC of the PV array, under
varying irradiance and temperature levels, has given rise to the fractional
VOC method .VMPP ≈ k1*VOC
where k1 is a constant of proportionality. since k1 is dependent on the
characteristics of the PV array being used, it usually has to be computed
beforehand by empirically determining VMPP and VOC for the specific PV
array at different irradiance and temperature levels. the factor k1 has been
reported to be between 0.71 and 0.78.
Once k1 is known, VMPP can be computed with VOC measured periodically
by momentarily shutting down the power converter. however, this incurs
some disadvantages, including temporary loss of power.
Once VMPP has been approximated, a closed-loop control on the array
power converter can be used to asymptotically reach this desired voltage.
the PV array technically never operates at the MPP. Depending on the
application of the PV system, this can sometimes be adequate. Even if
fractional VOC is not a true MPPT technique, it is very easy and cheap to
implement as it does not necessarily require DSP or microcontroller
control. However k1 is no more valid in the presence of partial shading
(which causes multiple local maxima) of the PV array and proposes sweeping
the PV array voltage to update k1. This obviously adds to the implementation
complexity and incurs more power loss.
61
OPEN VOLTAGE TEST
This technique refresh the voltage reference value every 3 s through the open
voltage measurement (for this measurement is necessary 10 ms without
power generation).
SIMULATION RESULTS
62
63
2.5) FRACTIONAL SHORT CIRCUIT CURRENT [5]
Fractional ISC results from the fact that, under varying atmospheric
conditions, IMPP is approximately linearly related to the ISC of the PV array
IMPP ≈ k2 *ISC where k2 is a proportionality constant. Just like in the
fractional VOC technique, k2 has to be determined according to the PV
array in use. The constant k2 is generally found to be between 0.78 and 0.92.
Measuring ISC during operation is problematic. An additional switch usually
has to be added to the power converter to periodically short the PV array so
that ISC can be measured using a current sensor. This increases the number
of components and cost. boost converter is used, where the switch in the
converter itself can be used to short the PV array.
SHORT CIRCUIT TEST
This technique refresh the reference current value every 3 s through the
short-circuit current measurement (for this measurement is necessary
10 ms without power generation).
64
SIMULATION RESULTS
65
66
CONCLUSION
1) Its is clear that open voltage and short circuit methods are
less effective than P& O and Incremental Conductance
method because of lossing power during updating The values
of open voltage or short circuit beside it is approximated
methods and not precise especially the partial shading
moments.
2) It is clear also that short circuit is the worst method because it
updates the duty cycle in the worst direction and it will be
more clear later through implemented simulation part.
67
2.6) FUZZY LOGIC CONTROL [9].[10],[11]
Microcontrollers have made using fuzzy logic control popular for MPPT over
the last decade. fuzzy logic controllers have the advantages of working with
imprecise inputs, not needing an accurate mathematical model, and handling
nonlinearity.
Fuzzy logic control generally consists of three stages:
1) FUZZIFICATION
2) RULE BASE TABLE
3) DEFUZZIFICATION
Photovoltaiv Block Diagram
During fuzzification numerical input variables are converted into linguistic
variables based on a membership function similar to following figure of
membership function .
68
In this case, five fuzzy levels are used: NB (negative big), NS (negative small),
ZE (zero), PS (positive small), and PB (positive big). And through other
examples seven fuzzy levels are used, probably for more accuracy. In the
following figure, a and b are based on the range of values of the numerical
variable. The membership function is sometimes made less symmetric to give
more importance to specific fuzzy levels .
The inputs to a MPPT fuzzy logic controller are usually an error E and a
change in error ΔE. The user has the flexibility of choosing how to compute E
and ΔE.
General example of Membership Function of inputs & outputs in FLC
69
The linguistic variables assigned to ΔD for the different combinations of E and
ΔE are based on the power converter being used and also on the knowledge
of the user. following table is based on a boost converter. for example, if the
operating point is far to the left of the MPP (Fig. 2), that is E is PB, and ΔE is
ZE, then we want to largely increase the duty ratio, that is ΔD should be
PB to reach the MPP.
General example of Base Rule Table
70
In the defuzzification stage, the fuzzy logic controller output is converted
from a linguistic variable to a numerical variable still using a membership
function. This provides an analog signal that will control the power converter
to the MPP.
MPPT fuzzy logic controllers have been shown to perform well under varying
atmospheric conditions. However, their effectiveness depends a lot on the
knowledge of the user or control engineer in choosing the right error
computation and coming up with the rule base table. Experimental results
show fast convergence to the MPP and minimal fluctuation about it.
Many techniques of computing error E and change in error ΔE had been tried
So through the three following method A ,B & C will be displayed the most
known ones while all of them have same meaning that where exactly we are
at present from P-V or P-I curve .
71
METHOD (A)
Firstly we will be hinting about specification silicon of solar panel of this
experiment
Measured V-I as well as P-I characteristics for one solar panel are shown for
two insolation levels (%86 & %59) of the full insolation level .
72
Now we start to follow the next steps :
A) Determination of error (E)
B) FUZZIFICATION & MEMBERSHIP FUNCTIONS
73
C) RULE BASE TABLE
74
D) DEFUZZIFICATION
The output of fuzzy controller is a fuzzy subset .as the actual system
requires a non fuzzy value of control .defuzzification is required
several methods of defuzzification are available .of these ,the Mean of
Maxima (MOM) and Center of Area (COA) methods are most
commonly used .the COA method is usually selected for control
application .
Therefore COA method is used for defuzzification in the proposed
MPPT tracker of this experiment .
75
Figure Shows Aggregation and Defuzzification Process
76
77
SIMULATION RESULTS
CASE I : WITHOUT MPPT CONTROLLER
CASE II : WITH MPPT CONTROLLER
OPERATING CONDITIONS HERE IS HIGH INSOLATION
(@1.30 PM -4TH JUNE 2002)
78
CASE I : WITHOUT MPPT CONTROLLER
CASE II : WITH MPPT CONTROLLER
OPERATING CONDITIONS HERE IS HIGH INSOLATION
(@9.30 AM -4TH JUNE 2002)
79
METHOD (B)
A) Here the mentality is different for computing the error (E)
80
P-V Curve at Different Irradiance
B) FUZZIFICATION & MEMBERSHIP FUNCTIONS
81
82
C) RULE BASE TABLE
83
SIMULATION RESULTS
84
Power ratio is evaluated in this case to 96%
85
METHOD (C)
Here also the process is the same as previous ones but the mentality
of computing the error (E) Is different .
And this part will be explained later by section of implemented
Simulation .
CONSTRUCTED FUZZY MPPT SYSTEM
86
2.7) NEURAL NETWORK [12],[13],[14]
Along with fuzzy logic controllers came another technique of implementing
MPPT—neural network , which are also well adapted for microcontrollers.
Neural networks commonly have three layers: input, hidden, and output
layers as shown in following figure. The number of nodes in each layer vary
and are user-dependent. The input variables can be PV array parameters like
VOC and ISC, atmospheric data like irradiance and temperature, or any
combination of these. the output is usually one or several reference signal(s)
like a duty cycle signal used to drive the power converter to operate at or
close to the MPP.
How close the operating point gets to the MPP depends on the algorithms
used by the hidden layer and how well the neural network has been trained.
The links between the nodes are all weighted. The link between nodes i and j
is labeled as having a weight of wij as in the following figure. To accurately
identify the MPP, the wij ’s have to be carefully determined through a
training process
87
The PV array is tested over months or years and the patterns between the
input(s) and output(s) of the neural network are recorded. since most PV
arrays have different characteristics, a neural network has to be specifically
trained for the PV array with which it will be used. The characteristics of a PV
array also change with time, implying that the neural network has to be
periodically trained to guarantee accurate MPPT.
Example of Neural Network
88
NEURAL NETWORK TEST BY BACK PROPAGATION
TECHNIQUE
The network is fully connected to all neurons in the hidden layer through a
weight . Also bias signal is coupled to all the neurons through a weight. All
three layers of neural network have a hyper tangent sigmoid transfer
function .
The algorithm used for training is back-propagation. The back-propagation
training algorithm needs only inputs and the desired output to adapt the
weight.
Figure Shows the Inputs & Outputs of NN
89
This method performs a gradient descent on the error surface which is a
function of the line current error. A gradient descent minimization can be
performed on the error function. Back propagation training is referred to as
supervised training.
The neural network was trained using MATLAB software.
Feed Forward or Back Propagation Technique
What happens in this technique simply that after we introduce inputs and
desired outputs the neural performs some iterations to compare the actual
output with the desired optimum output And if there is deviation the neural
will back from output to input again passing through all blocks of bias and
weight to adapt it so it can show much better output in the following
iterations .
90
THE PROPOSED MPPT SCHEME
In order to minimize the long-term system losses, it is required that
converter input current has very small ripple and conversion efficiency is very
high even at part load.
Therefore the installation of a boost type converter or Cuk converter will be
advised. In this experiment a step-up converter is used as MPPT. The block
diagram of the proposed MPPT scheme is shown in the following figure .
Simulink Model of NN
91
SIMULATION RESULTS
The neural network is trained using program MATLAB to identify maximum
current and maximum voltage of solar array from data of radiation
(insolation level) level, and temperature. The conventional solar-array model
was used to simulate a real PV panel.
The simulation of system use following parameters:
The output power obtained from the boost converter is equal to the maximum
power excluding the converter losses. The converter has operating at switching
frequency Fs =25 kHz has maximum power Pmref=7.8W at normal operating
condition temperature .
92
CONCLUSION
The back propagation neural network was modeled and simulated.
The simulation results have shown that training of back
propagation neural networks gives closer maximum power point.
as the developed model takes care about the variations of all the
parameters with respect to environmental conditions, it can be
used to predetermine the PV characteristics.
93
2.8) LOAD LINE [15]
Here is suggested a method to accelerate convergence time of
conventional MPPT algorithms without sacrificing accuracy in steady
state. The power converter is controlled so as to inherently reflect a
virtual load toward the PV array. The virtual load is optimized to intersect
with the output characteristics of the PV generator in the vicinity of the
MPP, even under varying irradiation conditions. This approach combines
the simplicity of algorithms with constant step size and the improved
performance of those with varying step size.
Previous Figure presents the I–V curves of a PV panel for different insolation
levels, where the MPP loci are marked. A linear virtual load line is also
sketched, which corresponds to V − rI − Vref = 0.
94
MPPT converter operation along this line (at the MPPT converter’s input) is
achieved by the simple control loop depicted in the following Figure, where
last equation is accomplished by a current sensor with appropriate gain r.
Since the MPP loci are not on a straight line, the value of Vref is tuned as
well (most likely in software), so that the virtual load line moves to different
locations, while maintaining its inclination that is set by the gain r. In this
way, the PV generator is operated at the actual MPP at any given insolation.
Tuning of Vref is accomplished iteratively, by either P&O or the incremental
conductance algorithm .
Proposed accelerated control scheme
95
Two convergence methods after step change in insolation,
Solid line: Proposed method
Dashed line : hill climbing method
Previous Fig describes the operation before and immediately after a step in
insolation for two cases: the case of simple duty ratio control (broken line) and
the operation due to proposed accelerated control (solid line). If the PV array
was operating at MPP “1,” and at an insolation of 0.3 Sun, then after the
insolation stepped to 0.6 Sun and just before the MPPT controller performed
the following iteration, the PV panel operates at point 2 (in case of accelerated
control) or 2 (in case of conventional hill climbing control). Eventually, after
several iterations, the operation points move to the actual new MPP “3” via an
iterative process. During the rapid change from “1” to “2,” the voltage and
duty cycle step sizes change (for the accelerated controller); however, it is not
being explicitly computed.
The step size is generated automatically by the control loop that implements
(1). Evidently, operation point 2 is much closer to the new MPP (point 3) than
to point 2. Therefore, less iteration would be required till the operation at the
actual MPP is attained.
96
SIMULATION RESULTS
Dynamic convergence due to a 100% step in insolation (0.3–0.6 Sun). Same
Its clear that the higher temperature the lower open circuit voltage will be
realized also the lower temperature the higher power will be extracted .
130
Effect of irradiance on PV power
Power & current are getting much higher with irradiance rise .
131
Effect of paralleled branches number on PV power
Here it is clear that the number of paralleled branches is the multiple factor
of PV power & current .
132
3.3) MPPT Methods SIMULATIONS BY GUI-
MATLAB
Through this part we will be displaying implemented simulations by GUI-
Matlab Which had been implemented by
Master Student : Ahmed Abd El Motaleb
& checked By
Tutor: Dr. Antonio de la Villa Jaen
Implemented methods in this chapter :
1) P & O
2) Modified P & O
3) Incremental Conductance
4) Open Circuit Voltage
5) Short Circuit Voltage
6) Fuzzy Logic
7) Neural Network
133
So here we will be starting with P & O method taking in consideration that
we make the step voltage fixed through the first two mentioned methods as
(0.5V) simply because By this step voltage we extract the optimum energy
from PV module with accepted ripple level but if the step voltage is lower
than 0.5 V then the ripples will be lower but the MPPT system will be slower
to reach maximum power so the extracted energy will be lower , and if the
step voltage is higher than 0.5 V then the ripples will be in too high level and
it will not be accepted by the converter and the average extracted energy will
be lower than case of step voltage =0.5V .
Please take in considerations that the following methods have the same
mentality of their counterparts which had been explained through last
chapter just fuzzy logic method have different mentality in this chapter will
be explained at its part .
134
Also we shall mention that all applied following methods had been
implemented under the following conditions :
*Energy Gap (Eg) = 1.12v
*Number of series cells Ns = 72
*Number of Paralleled branches = 5
*Temperature = 25 C
*Starting tracking voltage for ( p & o , Incremental Conductance ) methods
only = 26 V Starting Tracking voltage is the starting voltage which we will be
starting through it to track the MPPT points ,and of course we can change
this value even starting from (0) voltage but we chose it as 26 v because it is
the nearest point to all MPPT under different insolation conditions .
135
3.4) Conventional P & O
G-Time Curve is the irradiance curve
136
Section of the peak curve of POWER VS TIME shows the oscillations during
P& O method .
137
138
139
3.5) Modified P & O
140
141
CONCLUSION
Its clear that modified P& O method is much better than
conventional P & O one because simply it tracks MPPT points
faster when PV module is performing far from MPPT point and
slower when the PV is close to MPPT point so exactly modified
P & O is directly proportional with the slop of P-V curve
142
3.6) INCREMENTAL CONDUCTANCE
143
144
Now we will apply 4 th case for both conventional P & O method and
Incremental Conductance method so , the 4 th case is step up &
down of insolation but under very low insolation so we will have very
important conclusion will be shown after following simulations .
145
Conventional P & O
146
INCREMENTAL CONDUCTANCE
147
CONCLUSION
Through last 1 st , 2 nd & 3 rd cases of insolation , the insolations through this
cases are considered medium and high , and we can observe that the extracted
energy from each case In P & O method is lower than its counterpart in
Incremental Conductance one but through the 4 th case we can observe that
the extracted energy from both methods are the same even under different
step voltage values which will be also shown by the end of this Chapter .
And that because through low insolation cases the slop of P-V curve will be
approximately the same all over the time so incremental conductance will not
realize any difference in slope and for P & O just one step voltage change will
impress high power variation so both methods will not be precise under low
insolation level .
148
3.7) OPEN CIRCUIT VOLTAGE
Here we adopt same mentality of open circuit voltage which was explained in
the last chapter , also here we set the disconnecting time of open circuit to
be 10 milli-second and the refreshment time to update the open circuit
voltage value to be three seconds .
Figure explains the behaviour of open circuit voltage method that within one
sample of voltage refreshment if the insolation changes then this method will
not be able to track the MPPT of all insoltaions only the insolation of the
specified sample .
149
150
151
152
3.8) SHORT CIRCUIT CURRENT
Here also we adopt same mentality of short circuit current which was
explained in the last chapter , also here we set the disconnecting time of
short circuit to be 10 milli-second and the refreshment time to update the
short circuit current value to be three seconds .
Figure explains the behaviour of short circuit current method that within one
sample of current refreshment if the insolation changes then this method will
not be able to track the MPPT of all insoltaions only the insolation of the
specified sample .
153
154
155
156
Previous figures explain very well that open voltage is much better
than short circuit method , hence the difference between maximum
power in short circuit method and the red dashed line ( maximum
available Power which can be extracted from PV) is higher than
open circuit voltage method .
157
CONCLUSION
Open circuit voltage method is much better than short circuit
current one hence the extracted energy from first one is higher
than the second , and that means that short circuit method tracks
the MPPT in the worst direction .
However the previous two methods , still the P& O or Incremental
Conductance methods are much better than them , and that is
clear from energy comparison .
158
3.9) ARTIFICIAL INTELLIGENCE PART
Artificial intelligence consists of three methods :
1) Fuzzy Logic
2) Neural Network
3) Genetic Algorithm
The system which will include any method of the previous ones is
called intelligent because it will be more aware with the variations
of system much more than other methods of control .
We had implemented through this thesis
1) Fuzzy Logic
2) Neural Network
159
3.10) FUZZY LOGIC Control
Here we adopt same mentality of fuzzy logic process as what was explained
in the last chapter
The only difference is how we calculate here the error
Error = dP/dV
dP = P(t) – P(t-1)
dV = V(t) – V(T-1)
if dp/dv is positive it means we are on the left side
if dp/dv is negative side
dp/dv =0 then we are at MPPT POINT
d(dp/dv)/dt determines the direction and rate of the change of MPPT
tracking movements and iterations .
160
Simulink explains the process of fuzzy logic controller
161
Membership function of the first input (Error)
Membership function of the second input (Change in error)
162
Membership function of the output (Step Voltage)
Three dimension surface shows relation between the two inputs & the
output
163
Figure shows the (Mapping Process) as the two inputs will be mapped to the
output region after that the total area of the produced output will be summed
and the center of gravity will be determined .
164
Base Rule Table
NL NS ZERO PS PL
NL NL NL NL NL NL
NS NS NS NS NS NS
ZERO NS NS NS NS NS
PS PS PS PS PS PS
PL PL PL PL PL PL
ER
RO
R
CHANGE IN ERROR
165
NOTE :
The following curves shows relation between output power of fuzzy
controller corresponding to time it mainly shows the performance
of fuzzy controller
Red line: represents the output power from fuzzy controller
Blue line : represents the maximum available power that could be
extracted from PV
166
Applying the first case of irradiance as what happened through last methods
in this chapter
167
Applying the second case of irradiance as what happened through
last methods in this chapter
168
169
3.11) NEURAL NETWORK
What happens here that we introduce some of the chracteristics of PV to the
neural network and this characteristics are insolation & temperature and of
course the desired output which will be in this case the optimum power
which represents the MPPT points .
We simulate three dimension map for power by matlab and introduce it to
neural network then neural will proceed to back propagation method which
will enable it to produce output Power and of course from first iterations the
output power will not be the same as the optimum introduced one , so back
propagation process starts and the neural will always will check in the
backward direction all the weights and bias of neural layers to improve its
output step by step till we get the desired result .
170
Three dimension power map at all possible temperature and insolation , it
represents the optimum desired power
Other thing shall be mentioned more data we introduce to neural ,more
aware it will be to system surface and can realize exactly the optimum power
at any insolation or temperature .
171
Figure represents the hidden and output layer of the used neural in this
example
Figure represents the hidden layer transfer function
172
Figure represents the hidden layer which consists of 3 neurons included
weight and bias for each one
Figure represents the output layer transfer function
173
NOTE :
The following curves shows relation between output power of
Neural controller corresponding to time it mainly shows the
performance of Neural controller
Red line: represents the output power from Neural controller
Pink line : represents the maximum available power that could be
extracted from PV
174
Here also we start to apply the three cases of insolations as mentioned
through last methods .
175
176
177
Figure 1 shows that the output power from neural network (P applied)
coincide exactly on the optimum desired power (Pref)
Figure 2 shows the oscillations produced from P & O method that the output
power (blue line) oscillates and cannot reach the optimum power (red line)
Conclusions of this comparison to show that neural network is much better
that conventional methods such P & O .
178
Simulink explains the process of neural network controller
179
CONCLUSION
Its clear that artificial intelligence methods such as fuzzy logic or
neural is much better than conventional methods , faster to reach
MPPT points , less oscillations and track always in right directions .
However we can not determine which method of artificial
intelligence is better .
Simply fuzzy logic can be more precise by more adapting to its
membership fuction and adding other membership function such as
negative medium and positive medium (NM) & (PM) .
Also we shall know that neural shall be trained before uploading it
to microcontroller more than one time to show its best
performance because it will not show its best result from first
training or iterations .
180
3.12) FINAL CONCLUSION
* Through this study we had mentioned in brief benefits of PV
system .
*We had mentioned some applications of PV .
* We were also concentrating on MPPT problem and its system
components .
* MPPT has endless methods to be applied .
* However form our simulations we had reached the following
results :
*Disadvantages of P & O method and that incremental
conductance is more effective than it simply because Incremental
senses the variation in irradiance while It is not detected in P & O .
* Open circuit voltage method is much better than short circuit
current method and that because short circuit method changes the
step voltage in worst direction .
181
* From extracted energy results we had found out that both P & O ,
Incremental Conductance are more effective than open circuit
voltage and Short circuit current .
* However the best results had been obtained from artificial
intelligence part represented in both fuzzy logic & Neural
Network
* Approximately there is no superiority for fuzzy or neural as both
of them outputs the same results nearly depending on MMF &
Base Rule Table for fuzzy and trained data for Neural Network .
* Other rest methods had been approved and outputs good results
but it might not be obvious for the latter to choose which one
better suits their application needs.
The main aspects of the MPPT techniques to be taken into
consideration are highlighted in the following subsections :
182
1) Implementation :
The ease of implementation is an important factor in deciding which MPPT
technique to use. However, this greatly depends on the end-users’ knowledge.
Some might be more familiar with analog circuitry, in which case, fractional ISC
or VOC and RCC are good options. Others might be willing to work with digital
circuitry, even if that may require the use of software and programming. Then
their selection should include hill climbing/P&O, Inc Cond, fuzzy logic control
and neural network .
2) Sensors
The number of sensors required to implement MPPT also affects the decision
process. Most of the time, it is easier and more reliable to measure voltage
than current. Moreover, current sensors are usually expensive and bulky. This
might be inconvenient in systems that consist of several PV arrays with
separate MPP trackers. In such cases, it might be wise to use MPPT methods
that require only one sensor or that can estimate the current from the voltage.
183
3) Multiple Local Maxima
The occurrence of multiple local maxima due to partial shading of the PV
array(s) can be a real hindrance to the proper functioning of an MPP tracker.
Considerable power loss can
be incurred if a local maximum is tracked instead of the real MPP. As
mentioned previously, the Generation Control Circuit should track the true
MPP even in the presence of multiple local maxima. however, the other
methods require an additional initial stage to bypass the unwanted local
maxima and bring operation to close the real MPP .
4) Costs
It is hard to mention the monetary costs of every single MPPT technique unless
it is built and implemented. This is unfortunately out of the scope of this paper.
However, a good costs comparison can be made by knowing whether the
technique is analog or digital, whether it requires software and programming,
and the number of sensors. analog implementation is generally cheaper than
digital, which normally involves a microcontroller that needs to be
programmed. Eliminating current sensors considerably drops the costs .
184
Note : this table explains all requirements of most important MPPT techniques
so the designer can determine the most suitable technique and the lowest price
according to the application ,some of this methods had been explained through
our work but not all of the listed methods .
185
References
[1]A Modified MPPT Method with Variable Perturbation Step for Photovoltaic System Chao Zhang, Dean Zhao,Jinjing Wang,Guichang Chen
[2]A New MPPT Method for Photovoltaic Generation Systems Based on Hill Climbing
Algorithm Dimosthenis Peftitsis*, Georgios Adamidis* and Anastasios Balouktsis
[3]Improved Perturbation and Observation Method (IP80) of MPPT Control for Photovoltaic Power Systems Youngseok Jung’). Junghun So’), Gwonjong Yu’), Jaeho Choi’)
[4]Improved MPPT method for rapidly changing environmental conditions D. Sera, T. Kerekes, R. Teodorescu, and F. Blaabjerg
[5] Experimental test of seven widely-adopted MPPT algorithms M. Berrera, A. Dolara, Student Member, IEEE, R. Faranda, Member, IEEE,
and S. Leva, Member, IEEE.
[6]Study on Realizing MPPT by Improved Incremental Conductance Method with Variable
Step-size Zhou Yan, Liu Fei, Yin Jinjun, Duan Shanxu
[7]A Variable Step Size INC MPPT Method for PV Systems Fangrui Liu, Shanxu Duan, Fei Liu, Bangyin Liu, and Yong Kang
[8]Analysis and Improvement of Maximum Power Point Tracking Algorithm Based on
Incremental Conductance Method for Photovoltaic Array Bangyin Liu, Shanxu Duan, Fei Liu, and Pengwei Xu
[9]MPPT with Asymmetric Fuzzy Control for Photovoltaic System Chao Zhang, Dean Zhao
[10]Study on the Intelligent Fuzzy Control Method for MPPT in Photovoltaic Voltage Grid System Xiao-bo Li1 Ke Dong2 Hao Wu3
[11]Optimisation of The Photovoltaic Grid Connected System Using Fuzzy Logic Control N.TKOUTI, A.MOUSSI
[12]Solar Array Modeling and Simulation of MPPT using Neural Network R.Ramaprabha1, Member, IEEE, B.L.Mathur2 and M.Sharanya
[13]An Introduction to Neural Networks Vincent Cheung,Kevin Cannons
186
[14]Solar-Array Modelling and Maximum Power Point Tracking Using
Neural Networks S. Premrudeepreechacharn, Member, IEEE and N. Patanapirom
[15]A Modified MPPT Scheme for Accelerated Convergence Michael Sokolov and Doron Shmilovitz
[16]Design and Implementation of A Novel MPPT Controller Based on Sun Tracking
Technology Guanghui Li Xinchun Shi Chao Fu Guoliang Zhou
[17]Sun tracking by peak power positioning for photovoltaic concentrator arrays. Control
Systems Magazine, IEEE Volume 3 Pritchard, D
[18]Dynamic Maximum Power Point Tracker for Photovoltaic Applications
Pallab Midya Motorola ,Inc .Corporate Development Labs Schaumburg , Illions