artigo para ser traduzido para o inglês
TRANSCRIPT
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
1/13
ELECTRONIC PROCESSING OF PHOTOVOLTAIC SOLAR ENERGY IN SYSTEMS
CONNECTED TO THE POWER GRID
UniversidadeTecnolgicaFederaldoParan
Av.MonteiroLobato,km04,PontaGrossa,Paran-CEP:84.016-210
UniversidadeFederaldeSantaCatarina,INEP
Caixapostal5119,Florianpolis, SantaCatarina-CEP:88.040-970
ABSTRACT
This article presents a photovoltaic system connected to the
commercial power grid in a centralized position and built
with a three-phase inverter with two stages which are capable
of extracting the maximum potential of photovoltaic panels.
The P&O algorithm is adopted as a technique of MPPT. The
isolation is made by a high frequency transformer. The
converter that composes the CC-CC stage dispenses thecontrol structure because it works with a ciclical reason and a
constant frequency within all of its operating range. This
makes the use of resonant CC-CC converters a good option
whose output is elevated in high frequencies favoring the
compression of power stucture. This way, the three-phase
converter with a resonant series is chosen to be part of this
stage. The MPPT is transferred to the CC-CA stage, that has a
electric current controler to the current injected in the electric
grid. The PWM three-phase inverter fed in tension, that
composes the CC-CA stage is controled and modulated in a
vectorial manner. The vectorial control imposes the Park
transformation to the line currents, from which results thedirect axis current. The MPPT uses the same variables from
the electronic current controler and maximizes the direct axis
current that reflects the power extracted from the photovoltaic
arrangement. This way, no specific measuring for the MPPT
is made, resulting in an economy of sensors.
Artigosubmetidoem29/10/2008(Id.:00914)Revisadoem19/03/2009,05/08/2009
AceitosobrecomendaodoEditorAssociadoProf.DarizonAlvesdeAn-drade
ABSTRACT
ElectronicProcessingofthePhotovoltaicSolarEnergyInGridConnectedSystems
Thispaperpresentsagrid-connectedphotovoltaicsystemin
centralizedconfigurationandconstructedwithathree-phase
dual-stageinverterabletoextractthemaximumpowerof
thePVmodules.TheP&Oalgorithm isadoptedasMPPT
technique.
The
isolation
is
achieved
by
a
high-frequencytransformer.
TheconverterusedintheDC-DCstagedis-
pensescontrol-loop, itsduty-cycleandswitchingfrequency
areconstantsthroughoutthepoweroperationrange.Thisen-
ablestheapplicationofDC-DCresonantconverters,whose
efficiency ishighathigh frequencies,favoringcompaction
of thepower circuit.Thus,the three-phaseseries resonant
converterischosentointegratethisstage.TheMPPTis
transferredtotheDC-ACstage,which,invariably,hasagrid
currentcontroller.Thethree-phasePWMvoltage-source in-
verter,intheDC-ACstage,usesvectorcontrolandspace
vectormodulation.ThevectorcontrolrequiresaParks
transformation
from
the
grid-currents
which
yield
the
directaxis
current.
The
MPPT
uses
the
same
variables
of
the
grid
currentcontrollerandmaximizethedirectaxiscurrentwhich
reflectsthepowerfluxfromthePVarray.Thus,anyspecific
measurementtorealizetheMPPTisnotneeded,resultingin
asmallcountofsensors.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
2/13
KEYWORDS:
photovoltaic
system,
three-phase
dual-stageinverter,three-phaseseriesressonantconverter,MPPT.
1 INTRODUO
In this study a modified two-stage inverter is presented, which
is used in the processing of photovoltaic solar energy in
systems connected to the power grid. This is a innovative
equipment in many aspects whose conception emphasized the
reduction of costs, volume and weight. In this conception,
only commercially viable questions were addressed although
this application still isnt allowed by law in Brazil.
The direct conversion of solar energy to electricity is made by
photovoltaic modules. The cost of these equipments is the
main factor that defines the option for other generating
sources. A photovoltaic system does not produce toxic waste
as nuclear plants do, dont pollute the environment as
thermoelectric, coal or gas plants and dont involve any
environmental or social impact as do hydroelectric plants.
This way, the fundamental justification for the development
of this job are the great expectations for the reduction in the
cost of constructing solar cells, projecting something below
US$0,40/watt against the current US$4,00/watt
(Burger,2008). This reduction will proportionally lead to therise in the weight that the inverter exercises in total cost of the
system. Figure 1 shows a graph elaborated from data obtained
from the consultation to many inverter resellers for the
European and North-American markets, e.g. Alter Systems
(2008).
Seeing as the medium cost for the generation of
hydroelectricity is of US$1,00/watt, this same investment re-
Figure 1: Medium price announced of inverters connected to the power grid with
power up to 40kW.
relation per watt verified for an inverter of 10kW, the needfor a technological improvement of the inverters connectedto the grid becomes clear.
1.1 Brazilian Energetical Context
The system for production and transmission of electricity inBrazil is a hydrothermal system of great proportions with astrong predominance of hydroelectric plants that currentlysupport 80% of the electricity generated in Brazil.
To compensate the lack of investments in hydroelectric plants,the federal government created a program for theconstruction of gas powered thermoelectrics. In periods ofunfavorable hidrological conditions, these contribute to thecompliance to the consuming market in a complimentarymanner (Agency, 2005).
In the year of 2007, it was registered the greatest amount inthe consuming of electricity in the decade, about 5%. Thisgrowth is credited to sustainable growht of Brazil. Therationing, though, thanks to the low level of water reservoirswas evaded thanks to the activation of the thermoelectric
plants. But the largest part of the gas available in the marketis compromised with companies that, in the past few yearshave been using insume and with a growing fleet of vehiclesthat are moved by gas. Recent tests made by the Brazilian
National Operator of the Electrical System showed thatabout half of the capacity of the thermoeletric plants couldnot be reached due to the lack o fuel (O Racionamento,2007).
There will only be a sustainable development withinvestments in the generation of energy coming fromrenewable, in other words, those that dont consume fuelsand dont produce harmful wastes. This way, theconservation of the energy that is provided byhydroelectrical plants, role that is performed bythermoelectrical plants, could be auxiliated by photovoltaicsolar systems that are quite adequate to the integration withthe general urban environment.
Brazil has a high daily medium rate of solar radiation,reaching more than 5kWh/m per day in some regions(Agncia, 2005). By using photovoltaical modules with 40%efficiency that are in a initial phase of industrialization(Clula, 2007), 2kWh/m would be generated daily. This way,less than 10m on average, would be enough to supply oneconsuming unit of Santa Catarina whose average consuming
is of 503kWh per month, the largest of the south of thecountry (Celesc, 2008). This state has an average of 2,7inhabitants per consuming unit, in other words, permeasuring point.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
3/13
1.2 Photovoltaic Systems
Photovoltaic systems use tension inverters keyed to the
conditioning and synchronism of the exit of the
photovoltaic arrangement with the electrical grid. The
control exercises two main functions, tracking the point of
maximum power operation (MPPT) of the photovoltaic
arrangement and injects a sinusoidal current on the grid,
with a power factor close to the unitary.
There are three topologies of inverters that are connected to
the electrical grid (Carrasco et alii, 2006):
One stage inverters: in just one processing stage
the MPPT and the control of the current injected on
the grid are made.
Figure 2: Inverter with just one energy processing stage.
Two stage inverters: a CC-CC makes the MPPT
while a CC-CA converter is responsible for the
control of the current injected on the grid.
Figure 3: Inverter with two energy processing stages.
Multi-stage inverter: Many CC-CC converters
respond by the MPPT and one CC-CA converter
takes care of the current injected on the network.
Figure 4: Inverter with multiple energy processing stages.
The ways by which the photovoltaic modules are combined with the
topologies of inverters present themselves in four different
configurations.
Centralized inverter: Photovoltaic modules are
connected in series to form a row. Rows are connected
in parallel to form the arrangement, that provides
energy to the CC bus of na inverter. Seen in figure 5.
Parallel CC: Rows or arrangements are connected to
CC-CC converters. Just one internal CC bus feeds a
CC-CA converter. Seen in figure 6.
Parallel CA: Rows or arrangements are connected to
individual inverters. These inverters exits are
connected internally in parallel at the side of the CA.
Seen in figure 7.
Integrated inverter: each photovoltaic module has a
small inverter. These inverters are connected in parallel
at the side of the CA. Seen in figure 8.
Table 1 relates the topologies of inverters connected to the grid
with the possible configurations of photovoltaic systems.
1.3 Considerations
In some countries, like the U.S.A, the galvanic isolation between
the photovoltaic arrangement and the electrical power grid is
mandatory. Hence, thats why modern inverters tend to use a
high frequency transformer (Carrasco et alii, 2006). This
happens because the low frequency transformer is bigger,
heavier, more expensive and less efficient (Rashid, 2001). For a
high frequency transformer to be able to integrate the topology
of the inverter that is used in a photovoltaic system, the CC-CC
stage becomes indispensable.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
4/13
Table 1: Options of implementation of a photovoltaic system
Figure 5: Centralized inverter conected to a photovoltaic arrangement.
Figure 6: Modules and inverter in a parallel CC configuration.
Figure 7: Modules and inverter in a parallel CA configuration.
Figure 8: Inverters integrated to their respective photovoltaic
modules.
According to Carrasco et alii (2006), in his considerations about
future tendencies, for the reduction in the cost per watt of inverters
connected to the electrical grid we would need to adopt a
centralized configuration. Besides, this configuration is pointed as
being the most appropriate for powers above 10kW, for being of
high efficiency and of low specific cost.
hhh Topologieshhhh
Configurations
hhhhhhh
Inverter
of one
stage
Inverter
of two stages
Inverter
of multiple
stages
Centralized inverter X
Parallel CC X X
Parallel CA X
Integrated inverter X
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
5/13
From these considerations, you can direct the atention to the
two-stage inverter with a high frequency isolation which is
configured in a centralized plant of medium power. According
to Carletti et alii (2005), a medium power plant will operate
between 10kW to 500kW. A study about the structures that
could be employed on the two stages of the inverter will be
made next.
Carrasco et alii(2006) presents the Full-Bridge converter as the
most well employed representative of the CC-CC stage in
powers above 750W thanks to its good factor of transformer
usage. But in Ziogas et alii(1988), it is demonstrated that the
three-phase inversion is advantegeous when compared to a
monophasic one. Generally speaking, it is possible to see a
better distribution of losses in three-phase structures such as
that you can see that the components are forced to withstand a
relatively smaller effort. Capacitive or inductive filters have
their volume sensitively reduced because the current that goes
through them has a frequency which is higher compared to a
monophasic converter. Speaking specifically of the CC-CC
stage, Ziogas et alii (1988) demonstrated a reduction in the
volume of the high frequency three-phase inverter if compared
to a Full-Bridge. The combination of these advantages leads to
a improvement in the dynamic behavior, with faster answers.
Table 2 was elaborated by consulting the catalogs of some of
the main representatives of the market of inverters which are
connected to grid in the world. Its purpose is to maintain a
close relationship between this work and the commercial line.
All inverters are three-phase ones beyond 8kW. Non-isolated
equipments are the majority up to 10kW. They are light and
compact. From 5 to 10kW the parallel CC configuration is
preffered. With two to three CC-CC converters at the entrance,
there is more flexibility in the assembly of the photovoltaic
arrangement. Beyond 10kW all structures tend to be isolated.
The low frequency transformers are the most common ones,
which are installed in bulky and heavy inverters. The parallel
CA configuration is quite adopted to allow the use of high
frequency transformers. Some manufacturers even use fifteenCC-CA converters in a parallel CA configuration. The
centralized configuration with a high frequency isolation is the
combination that conducts electricity to the lighter equipments.
The MPPT is made at the CC-CC stage in all inverters that are
connected to the grid and that have a CC-CC stage, specially those
that are high frequency isolated.
The 15kW output seems to be the most promissing one in Brazil.
Even though the cost/watt tends to reduce, there are two practicallimitations for higher outputs. There could be a difficulty to find
areas that are big enough for the instalation of the photovoltaic
modules (an estimated amount of 130 of area would be needed for
15kW panels of our current technology) and not all secondary
three-phase transformers would accept the inverters connection
(generally, the smallest of the three-phase transformers are of
15kVA). It is is worth pointing that the inverter proposed in this
study is destined to urban areas, integrated to constructions. The use
of rooftops and windows would be an interesting starting point.
As conclusion, this study seeks to propose a two-stage inverter with
a high frequency isolation, configured in a centralized plant,
adequate to the processing of 15kW of outpot and that employs
three-phase inverters both in the CC-CC stage, and the CC-CA
stage.
2 MODIFIED TWO STAGE INVERTER
Although it is possible to consider the development of what was
proposed to this point as a contribution, questions related to control
still werent addressed. In this theme, this study proposes solutions
that give value to the inverter related to the state of the art.
In first place, one should keep in mind that both the MPPT and the
injected current controler of the grid are implemented in a digital
signal processor, which favors the usage of the vectorial control and
vectorial modulation. The realization of the MPPT frequently
passes by one or two fundamental methods which are known as
disturbance and observation (D&O) and incremental conductance
(IncCond). Of these, D&O are preffered specially in systems that
are are connected to the network and of high output (Hua et alii,
1998; Jing et alii, 2005).
In this study, it is proposed that the D&O is executed at the CC-CA
stage, in a way to make use of the processing capacity involved in
the control of the injected current in the electrical network, giving
origin to the Modified Two Stage Inverter, seen in figure 9.
This modified two stage inverter uses the same variables used in the
control of the injected current in the network to execute the D&O,
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
6/13
2
Ld
provoking a sensible economy of sensors. This way, with no
specific measurings for the D&O, it is possible to maximize the
output of the inverter and force the photovoltaic arrangement to
operate in the maximum power point, MPP.
The strategy for this to work is fundamented in a characteristic
that can be observed in CC-CC converters that makes them
reproduce in their exit terminals the same behavior of the
photovoltaic arrangement, when they are connected to it. A
very interesting aspect is that for the CC-CC converter to have
a behavior which is correspondent to the photovoltaic
arrangement, it must operate with constant ciclical reasons and
frequencies. It is demonstrated that this aspect makes
converters, in unfeasible aspects, to work with extreme
efficiency by all the operating range of the MPPT.
Figure 9: Two stage inverter with the MPPT being done on the CC-CA
stage.
2.1 CC-CC Stage
The most attractive three-phase CC-CC converters are those
that present smooth comutation since they can operate with
high keying frequencies which leads to significant reduction in
capacitive and magnetic elements. There arent many
three-phase CC-CC converters with smooth comutation
available in literature. A few of them (Doncker et alii, 1911;
Prasad et alii, 1991; Bhat e Zheng, 1996, Oliveira Jnior e
Barbi, 2005; Jacobs et alii,2004) were evalutated concerningtheir efficiency, number of components, emission of
electromagnetic interference, performance in assimetrical
operating conditions and aptitude to elevated outputs. The
structure presented in figure 10, obtained from Jacobs et alii
(2004), has showed itself as being insuperable in all aspects.
The transformer was echanged for its dispersion indutancies.
The transistors of one arm are activated in a complimentary
way. Between each arm there is a discrepancy of 120 in the
command pulses. When the keying frequency, f1, equals to the
resonance, fr, the converter operates in ZCS. If f1 > fr, the
converter will operate in ZVS. In this operating mode, a
reduction in the entry power, which is a normal opertating
condition of the arrangement, the converters efficiency reduces
drastically (Jacobs et alii, 2004). This way, for the ZCS mode, it is
possible to produce (1) and (2).
f
1
=fr=
1
In which: Iinav, VinavOutput and medium tension in the entry of
the CC-CC converter
IpaCurrent in the exit terminals of the photovoltaic arrangement.
Vdcav Medium tension in the exit of the CC-CC converter,
refered to the primary of the transformer.
RlossRepresents all the losses on the SRC3.
Although it has many advantages, the SRC3 has one deficiency
which makes it unfit to do the MPPT. Its keying frequency and its
ciclic reason are fixed variable (Jacobs et alii, 2004). So
concentrating all the control grids on the CC-CA stage is one of the
greatest merits of this study. This technique allows for such
defieciencies to be neglected.
2.2 CC-CA Stage
Photovoltaic arrangements have a behavior that is very similar to
that of a source of power. Hence why most of the CC-CA converters
that are connected to a network are fed in tension (Rashid, 2001), in
other words, have a CC bus in their entry, as shown in figure 11. The
high tension in the bus generally imposes the use of IGBTs in the
composition of the bridge of transistors.
The vectorial modulation, SVM, is employed in the deployment oftransistors in a way to minimize the THD of the current of the line.
A simetrical sequence is used to generate these trigger pulses. With
this, significant harmonic components only appear from keying
frequencies. The 5% limit of the THD which was established by
international laws is atended with the L filter of first order in the
connection with the electric grid. The combination of the vectorial
modulation with the use of the L filter of first order ended up with a
greater compactation and a reduced costs since the line inducers
were built with cores of iron and silicon. The ferrite is more
expensive and has a low density of magnetic flow.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
7/13
From figure 11, it is possible to see that the Park transformation
(vectorial control) is used in the modeling of the converter,
from which results the direct axis current, Id.
3 CONTROL STRATEGY
The direct axis current Id represents the medium output which
has been injected in the electric grid, which is maximized by
the D&O algorithm. In it, the tension the bus CC, Vdc, is
disturbed while theId current is observed. The logic of the
D&O algorithm is described in table 3, whose mathematical
representation is in (3).
Vdc(k + 1)= Vdc(k) +
V
sign
[Vdc
(k)
Vdc
(k
1)]sign
[Id
(k)
Id
(k
1)]
(3)
The V greatness corresponds to the amplitude of the disturbance
which is applied in Vdc. This disturbance, as well as the period of its
application are adequately dimensioned in a way to evade the
instability of control (Fermia et alii, 2005)
Thesignfunction, that appears in (3), only extracts the signal of the
calculation made in its argument.
The system of figure 11 has more than one control entry, which led to
the abandoning of the theory of classic control favoring the theory of
control based in the space of states. Thus, the project of a servossystem
with state feedback and full control was developed (Ogata, 2003). The
state variables are generated from only two current sensors and one
tension sensor.. For the MPPT, no additional measuring is needed.
The dynamic model of the converter from figure 11 is presented in (4)
in a way to give better evidence to the state variables.
--It was not possible to import (4)
Id, Iq, VdcState variables.
Iquaxis current in quadrature.
Dd, DqControl entries.
VgTension of efficient phase of the electric grid.
WAngular frequency of the electric grid.
L, RIndutance and intrinsic resistance of the line inducers.
C2Capacitor of the CC bus.
Idc Resulting current from the division between the output
delivered to the CC bus and its tension, Vdc.
4 PROJECT PROCEDURE
Figure 12 presents the output structure of the modified two stage
inverter positioned between the photovoltaic arrangement and the
three-phase electric grid.
Although it has been established that the nominal output of 15kW is
the most adequate for the brazilian energetic context, a prototype of
4kW is projected in function of the availability of photovoltaic
modules in laboratory. The photovoltaic arrangement is composed
of 20 modules (two lines) KC200GT from Kyocera, in a total of
4kWp. The kWp unit represents the maximum potential that can be
extrracted from the photovoltaic arrangement, that is, the output in
STC (Standard Test Conditions Solar radiation of 1kW/m,
spectrum of 1,5 of air mass and temperature of photovoltaic cells at
25C). The specifications for the project of the inverter are:
Pin=4kW ; Vinav = 263V ; Iinav = 15,21 ; f1= 40kHz ; f2
=20kHz ; Vin=1%;Vg=220V
Perturbao
emVdcVariaodeId
Prxima
perturbao
+V Crescente +V
+V Decrescente -V
-
V Crescente -
V
-V Decrescente +V
Figure 10:Resonant three-phase CC-CC converter (SRC3)
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
8/13
215
1
V
In which:
PinNominal entry output of the CC-CC stage.
F2Keying frequency of the CC-CA stage
Vin - Tension ondulation in the photovoltaic
arragements terminal.
The SRC3s efficiency is estimated to be in n1=0,97 (Jacobs et
alii, 2004). It is then possible to calculate the output which was
delivered to the CC bus, as well as its tension, as seen in (5) and
(6). Vdcav is elevated to 816V due to the turn ratio of the
transformer.
P
dc
=
1
P
in
=
3880W
(5)
Vdcav
= 1Vinav= 255V (6)
The capacitor docked to the photovoltaic arrangement is
calculated in (7). A polyester capacitor of 680nF was used.
C1=Iinav
From (8) it is possible to obatin the value of the capacitor of the CC
bus, C2. The undulation over this capacitor,VC2=0,2V,
corresponds to a percentage of V. The value of C2 = 333uF is
adopted. This is the commercial value resulting from the association of
three electrolytic capactior of 1000uF. It is worth noticing that the tension in
this bus can approach 1000V.
C2 P dc / (12
2
Vg
VC2
f
2)
From three-phase transformer, built from three monophasics, a dispersion
of value Ld = 1,64uH was resulted. This way, each resonant capacitor
equals Cr= 9,9uF. This value, calculated in (1) is obtained by the parallel
association of three capacitors of polypropylene of 3,3uF.
The line inducers are calculated by (9), resulting in L = 9,3mH. The
undulation in the line of current is worthIL=0,42A which
corresponds to 5% of the current in the line of peak.
2
VgL = 4
f
2
IL
5 THEORETICAL RESULTS
The losses in the CC-CC stage are calculated in 0,32 isolatingRlossin (2).
By exchanging Rloss and (6) in (2), it is possible to trace the entry
characteristics of the SRC3, shown in figure 13. The parameters that
define this feature are its inclination and position. The inclination depends
on the losses. The position, of Vdc. This tension is controlled by the
Figure 11: Three-phase CC-CA converter fed in tension with the MPPT variables
Figure 12: Modified two stage inverter.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
9/13
CC-CA converter. This way, the displacement in the I-V characteristic
of entry of the CC-CC converter is made by the CC-CA converter.
Figure 14 shows that the crossings between the curves that are
characteristics of the CC-CC converter and that of the photovoltaic
arrangement of 4kWp occurs practically over the points of maximum
output for various levels of radiation.
If the temperature varies, the CC-CA converter, by the P&O
algorithm, repositions characteristic I-V of SRC3 until it extracts the
maximum output from the photovoltaic arrangement once again. The
Id current reflects the output extracted from the arrangement. This
control action is ilustrated in figure 15. The tension Vdc is disturbed
in 4V every 50ms. Figure 16 shows the behavior of the exit output of
the photovoltaic arrangement by adjustments in the Vdc tension,
presented in figure 15. The error of relative tracking, R, will almost
reach zero in a permanent regime. This occurs thanks to the
aproximation between characteristic I-V of the entry of the SRC3 and
the MPP of the photovoltaic arrangement.
Figure 13: Characteristic I-V of entry of the SRC3 in nominal
conditions.
Figure 14: Crossing between the characteristics of the CC-CC
converter and the photovoltaic arrangement.
6 EXPERIMENTAL RESULTS
The three-phase CC-CA PWM converter fed in tension, adopted in
this study is the most used converter in the world when it is
necessary to inject energy that comes from a photovoltaic
arrangement in the three-phase electrical grid (Burget et alii, 2008).
Amongst the works that use this converter for these ends, some are
national (Schonardie e Martins, 2007; Cavalcanti et alii, 2006).
There are studies that point to advantages in its substitutions for
other structures like the three-phase CC-CA PWM converter fed in
current (Sahan et alii, 2008) or for multilevel inverters (Selvaraj and
Rahim, 2009). The most varied of current control in the power grid
techniques are employed and of MPPT. The LCL filter on the
interface with the network can make its structure more compact
(Blaabjerg et alii,2004). This is a very productive option for
researchers of this subjetct. The aplication for the three-phase
resonant series CC-CC converter in a two stage inverter is the focus
and the main contribution of this study. Thus, this section is
dedicated to the presentation of experimental results that relate to
the CC-CC stage of the two stage inverter.
Figure 17 presents the 4kW prototype implemented in laboratory.
Figure 15: P&O Algorith: Disturbance in Vdc as the variation
observed in Id.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
10/13
The SK20GD065 and SK10GD123 modules were used, both
from Semikron. Each module has six IGBTs of 600V and
1200V respectively. Two SKHI61 command circuits were used,
also from Semikron, take the trigger pulses to these transistors.
The rectifying bridge was implemented with six ultra fast
FFPF05U120S diodes of 1200V and 5. The conditioning board
has two LA 25-NP current sensors and one LV 25-P tension
sensor, both from LEM, so it is possible to monitor the status
variables. In it the synchronism signal is also generated, which
is needed for the controler. The DSP TMS320F2812 responds
by control of the current injected on the grid, by the MPPT and
by the protection of lack of phase, that stops the inverter from
continuing operation when the electrical grid is de-energized.
Simultaneously, the DSP activates the CC-CC stage by using
six PWM ports. Trigger pulses with a frequency of 40kHz are
generated with a duty cicle of 50% and a idle time of 640ns.
Figure 18 presents the resonant currents in nominal operation
conditions. These currents flow by the three-phase transformer,
comprised by three monophasic transformers. The induction
dispersion of each phase are not exactly equal, obviously. This
provokes small differences in the amplitudes of the resonant
currents, which doesnt harm the SRC3s operation and doesnt lead
to a risk of saturation to the transformer. In normal operatingconditions, the continuous tensions that could saturate the
transformer are blocked by the resonant capacitors.
Figures 19 and 20 present the entry and exit currents of the SRC3,
behind their respective filters. These currents have a frequency that
is six times greater than that of the keying and low undulation,
resulting in a continuous flow of output. These are not common
features of the CC-CC converters, especially the monophasic ones.
From these features the atypical value of 680nF for the entry
capacitor, C1.
Figures 20 and 21 favors the visualization of the asymmetry
provoked in the Iin current and in the Vin tension by the differences
in the dispersions inductancies. The use of commercial transformers
would certainly reduce such asymetry.
The ZCS commutation is shown in figure 22. In low outputs this
commutaion becomes dissipative, as shown in figure 23. Theefficiency, shown in figure 24 was around 97,5% in nominal
conditions. Thus, the losses in the SRC3 can be recalculated by (6)
and (2), resulting inRloss= 0,26.
By changing the newRlossvalue in (10), equation derived from (2), it is
possible to obtain the inclination of the I-V entry feature of the SRC3 for
different tension values of the CC bus. Figure 25 confirms this result.
The entry feature of the SRC3 favors the MPPT. This becomes
evident when figure 25 (Vdc = 816) is superimposed over the
characteristic curves of the photovoltaic arrangement for a certain
temperature, resulting in figure 26. For each value of solar radiation
a crossing point between the features of the arrangement and
converter are established. It happens that the inclination defined in
(10) makes these crossings much closer to the MPPs, dispensing the
action of the MPPT in the case of rapid variations of atmospheric
conditions. The temperature varies slowly.
The normal tension of the CC bus corresponds to 816V and its
minimum operation value is around 600V. The inverter needs
tensions above this value to be able to inject current in electric grid.
Figure 16: MPPT performance, analised in the exit terminals
of the photovoltaic arrangement.
Figure 17: The prototype of the modified two stage
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
11/13
Figure 18: Resonant currents in nominal conditions.
Figure 19: Entry and exit currents of the CC-CC stage and
tension between the collector and emitter of the S1
transistor in nominal conditions.
Figure 20: Entry and exit currents of the CC-CC stage and
tension between the collector and emitter of the S1
transistor for P dc=500W.
Figure 21: Tension in the exit terminals of the photovoltaic
arrangement for P dc = 500W.
Figure 22: Tension and current in the S1 transistor in
nominal conditions.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
12/13
7 CONCLUSION
In this study a conceptual change in the topology of the two stage
inverter employed on the processing of photovoltaic solar energy
connected to the power grid was made. Such conceptual change,
that consists in concentrating all of the control structure in the
CC-CA stage gives birth to the modified two stage inverter. In it,
no control action takes place over the CC-CC stage, in other
words, the frequency and ciclical reason are constant by all of the
operating range. Therefore, it was possible to apply the
three-phase inverter in its composition. Between many of the
advantages offered by this converter, the maximum undulation of
1% in exit tension of the arrangement, obtained with a polyester
capacitor of 680nF, whose cost is meaningless, excels.
Commercially, manufacturers adopt undulations of 2% below
10kW in outputs above 100kW and up to 10% for outputs below
10kW. These values usually are reached by capacitative banks that
reach some mili farads.
The low number of sensors that are required by the modified two
stage inverter is even more reduced by the innovative way in which
the D&O algorithm is executed, disturbing the tension on the CC
bus and observing the variation of the direct axis current. These
variables were generated as part of the control of the current
injected on the electric grid. This means that no specific measuring
for the MPPT was made.
Figure 23: Tension and current in the S1 transistor for P
dc=500W.
Figure 24: Efficiency of the SRC3.
Figure 25: I-V characteristic of entry of the CC-CC
converter for different tensions of the CC bus.
Figure 26: Superimposition of the characteristic curves of
the photovoltaic arrangements and of the entry of the
SRC3 for Vdc = 816V.
-
8/11/2019 Artigo Para Ser Traduzido Para o Ingls
13/13
The MPPT algorithms tend to become slower as they become
more precise. Rapid changes in the atmospheric conditions
occur frequently. So, abrupt variations in the solar radiation
generally cause important losses. In this inverter, the entry
feature of the three-phase converter compensates almost
entirely the MPPs displacements. Theacting time is limited by
the response time of the CC-CC converter, which is much fasterthan the MPPT grid.
ACKNOWLEDGEMENTS
The authors thank FINEP and the CNPq for the financial
support destined for the development of this study.
REFERENCES
Agncia
Nacional
de
Energia
Eltrica
(2005).
Atlas
de
Ener-giaEltricadoBrasil,2.ed.,Braslia.
Alter Systems: Alternative energy products and
services. Berkeley, CA. Disponvel em:
. Acesso em: 16
ago.2008.
Bhat,A.K.S.andL.Zheng(1996).AnalysisandDesignof
aThree-PhaseLCCTypeResonantConverter.IEEE
PowerElectronicsSpecialistsConference,Baveno,pp.
252-258.
Blaabjerg,F.,R.Teodorescu,Z.ChenandM.Lissere(2004).
Power
Converters
and
Control
of
Renewable
Energy
System.Proc.ofICPE04,pp.I2-I20.
Burger,A.K.NextGenerationDyeSensitive&Or-
ganic PV, Part 1. RenewableEnergyWorld.com,
Cadiz, Spain, 24 jul. 2008. Disponvel em:
.Acessoem:
16ago.2008.
Burger,B.,B.Goeldi,D.KranzerandH.Schmidt(2008).98.8%InverterEfficiencywithSICTransistor.23rdEu-
ropeanPhotovoltaicSolarEnergyConference,Valen-
cia,pp.2688-2692.
Carletti,R.L.,L.C.G. LopesandP.G. Barbosa(2005).Ac-
tive&ReactivePowersControlSchemeforaGrid-
ConnectedPhotovoltaicGenerationSystemBasedon
VSIwithSelectiveHarmonicElimination.Brazilian
PowerElectronicsConference,Recife,pp.129-134.
Carrasco,
J.M.,
L.G.
Franquelo,
J.T.
Bialaziewicz,
E.
Gal-
vn,R.C.P.Guisado,M.A.M.Prats,J.I.LenandN.M.
Alfonso(2006).Power-ElectronicSystemsfortheGrid
IntegrationofRenewableEnergySources:ASurvey.
IEEETransactionsonIndustrialElectronics,v.53,n.
4,pp.1002-1016.
Cavalcanti,M.C.,G.M.S.Azevedo,B.A.Amaral,F.A.S.Ne-
ves,D.C.MoreiraandK.C.Oliveira(2006).AGrid
ConnectedPhotovoltaicGenerationSystemwithCom-
pensationofCurrentHarmonicsandVoltageSags.Ele-
trnicadePotncia,v.11,n.2,pp.93-101.
Celesc:
Centrais
Eltricas
de
Santa
Catarina.
Perfil.
Dis-
ponvel
em: .
Acessoem16ago.2008.
Clulasolarbaterecordedeeficinciaepodeviabilizar
energiasolar.RedaodoSiteInovaoTecno-
lgica.Campinas,SP,26jul.2007.Disponvel
em: .Acessoem:16ago.2008.
Consumo
bate
recorde
e
o
maior
da
dcada.
InvestNews Tempo Real (Brasil), So
Paulo, 23nov.2007.Energia.Disponvelem:
.Acessoem:16ago.
2008.
Doncker,R.W.A.A.de,D.M.DivanandM.H.A.Khe-
raluwala(1991).Three-PhaseSoft-SwitchedHigh-
Power-Densitydc/dcConverterforHigh-PowerAppli-
cations.
IEEE
Transactions
on
Industry
Applications,
v.27,n.1,pp.63-73.
Femia,N.,G.Petrone,G.SpagnuoloandM.Vitelli(2005).
OptimizationofPerturbandObserveMaximumPower
PointTrackingMethod.IEEETransactions onPower
Electronics,v.20,n.4,pp.963-973.