Cascaded H-bridge multilevel converter for grid ... - IEEE Xplore

0 downloads 0 Views 390KB Size Report
Generators with Independent Maximum Power Point Tracking of each Solar Array. 0. Alonso, P. ... multilevel converter for grid-connected photovoltaic systems.

Cascaded H-Bridge Multilevel Converter for Grid Connected Photovoltaic Generators with Independent Maximum Power Point Tracking of each Solar Array 0.Alonso, P. Sanchis, E. Gubia and L. Marroyo Department of Electrical and Electronic Engineering Universidad Publica de Navarra 3 1006 Campus Arrosadia - Pamplona - Spain email: [email protected] Abstract - This paper introduces a new control method and proportional PWM modulation of the cascaded H-bridge multilevel converter for grid-connected photovoltaic systems. This control makes each H-bridge module supply different power levels, allowing therefore for each module an independent maximum power point tracking of the corresponding photovoltaic array.

1. INTRODUCTION Nowadays, grid-connected photovoltaic systems are the higher developing solar energy applications. In these systems, the use of all the available energy depends on the static converter topology that is used. It has been demonstrated that topologies that use a lower amount of panels for each array improve the global efficiency of the photovoltaic generator. This is a result of a reduction in mismatch losses, partial shadows of the array, etc. Obviously, the efficient use of the available solar energy of each group of panels requires the implementation of an independent maximum power point tracking technique (MPPT) for each one of these groups. There are several modular structures recently developed that work under the previous criteria. Many of these structures, called string converters, use a common DC bus where each one of the solar receiving modules transfers the power [l] [2]. Usually, these modules consist of a solar array and a DC-toDC converter controlled by a MPPT algorithm. Afterwards, the available energy on the DC bus is transferred to the grid

by means of a half-bridge or full-bridge converter. Fig. 1 shows an example of one of these structures [I]. There are other structures that use the multilevel modular configuration and whose circuits and basic characteristics are described in [3]. Among these structures, one that has similar features in modularity and control with respect to the previous string converter is the cascaded H-bridge multilevel converter, whose generalised power circuit is shown in Fig. 2. This converter takes advantages of one of the most important feature of the multilevel converters, that is, an output voltage synthesis with a higher number of levels. This number of levels grows according to the number of series-connected modules in the converter. This characteristic has some advantages. In one hand, a reduction in the common-mode perturbations is verified, that is originated by the leakage photovoltaic array earth capacitance. In the other hand, lower amplitude harmonics are injected to the grid and with higher frequencies, simplifying their filtering. Although these multilevel converters are initially prepared for applications where each one of the modules delivers the same power, their optimal use in photovoltaic generators require an independent delivering of each one of these modules. Therefore, in this paper a new control methodology and proportional PWM modulator are presented, that allow an independent MPPT implementation for each solar array. Simulation results have been carried out to validate the proposed control technique.

oTn ,,...

c3

ibu

r P Y h

o"*?run

: :

:

IJ

A

J

:

1

3

e*-K!%z--.j

Fig. 1. Generic single-phase String Converter

0-7803-7754-0/03/$17.00 02003 IEEE

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

j

73 1

11. BRIEF SYSTEM INTRODUCTION AND DESCRIPTION

Fig. 2 shows a H-bridge multilevel converter circuit generalised for n series-connected modules. By means of modulation techniques and keeping the same DC voltage for each module, the converter synthesises an output voltage V, with a number of levels equal to 1+2n. As a consequence, common-mode perturbations and voltage harmonic amplitudes are reduced, and therefore, the filtering processes results to be easier. Here, a bipolar PWM modulation method for each H-bridge module has been used. Under this modulation, the multilevel voltage synthesis is achieved by means of a phase shift displacement of the carrier waves of the different modules. This angle is calculated as follows: 3 60 ATc =n

With this modulation, a complete cancellation of output voltage harmonic groups is achieved. In fact, in ideal conditions where each module delivers the same power from the same DC voltage, the central frequency of the first harmonic group of the output voltage is: where fc is the switching frequency of any H-bridge module. When the previous conditions are not kept, the harmonic cancellation is not complete, and the output voltage has a worse harmonic spectrum. The simulation results show both operation possibilities and their effects on the output voltage and current. Fig. 3 shows a simplified block diagram of the whole control methodology that it is proposed. The first stage corresponds with the independent MPPT algorithms and capacitor voltage controls for each photovoltaic array. By means of using the averages values of their voltages and currents, the capacitor voltages are controlled with the aim of getting the maximum available power of each solar array. As

a result of these controls, the new medium current references (IDck) of each module are obtained. The next stage calculates the available power of each module; therefore, the sum of all of them (PT) will be the total available power to be transferred to the grid. This is done by means of a power control in the next stage, in which the output voltage VHTthat can be synthesised by the converter is calculated. In the last stage, the proportional PWM modulator shares the voltage Vm among the different H-bridge modules. As it will be demonstrated, the voltage fraction of a module is directly proportional to the power fraction that the module is transferring. 111. MPPTALGORITHM IMPLEMENTATION

There are several techniques to fulfil the detection and tracking of the maximum power point in a photovoltaic generator. In this paper a Perturbation and Observation (P&O) algorithm has been used, whose operating principle is exposed in [4]. This algorithm is implemented in digital processor and uses the average values of voltage and current of the solar array. Due to the active power delivering of each module to the grid, in the DC side of each converter there is a 100 Hz ripple in all the magnitudes. Therefore, the solar array current and voltage measures are filtered by means of a digital 100 Hz window filter. The different MPPT algorithms provide the average voltage reference of the corresponding photovoltaic generator. Every one of these references is processed by means of an independent average voltage control of the DC link capacitor for each H-bridge module. Fig. 4 shows the block diagram of a generalised voltage control for any k-th module. From each one of these stages the average reference current IDCk are obtained. These magnitudes and the corresponding digital filtered voltages VDCk are sent to the next stage where the power control is camed out.

732

k ~ - P Varray h4PPT VDCk

vDck.REF

cs IPVI

IPVk

Converrer

~,

; I

IOOHz Window Digital Filter

I

7,

ipw

IOOHz Window

Fig. 4. Capacitor voltage control loop of each module

Fig. 5. Simplified output circuit and unitary factor power operation

I vgrid (nns

value)

v. PROPORTIONAL PWM MODULATOR

IV. POWERCONTROL The total available power is the sum of the calculated power of each solar array. That is: PT = 4 + P2+ ..+ P" = VDCIIDCl + VDC21,, + ..+ VD,IDC"

(3)

Due to the fact that the power is transferred to a more or less constant voltage grid, the power control can be done in an indirect way by means of the output current (IL) control. As it can be observed in Fig. 5 , the power factor will be 1, and for that the rms-value of the output reference current is calculated as follows: (4)

Fig. 6 shows the block diagram of the power control by means of the output current control, whose output is the output voltage V, that must be synthesised by the converter. For this control, the modulator and the converter stages are considered a block with a unitary gain.

The main task of the modulator is to synthesise the output reference voltage V., Due to the series-connection of the modules it is verified that:

Vm = VNI+ v,,

+ ...+ v,,

(5)

The proportional relationship between each individual voltage and the total voltage Vm defines what has been called proportionality factors:

In this converter, each module drives the same output current I,. Therefore, each voltage module is proportional to its power, verifying the following relationship:

733

factors that have been obtained. As it can be observed, during the initial stage the power delivered for each module is 680 W, what implies an output current with a maximum value of 13.2 A, and a proportional factors equal to a k= 113 . After the change of irradiance, and in steady-state conditions, it can be demonstrated how the proportional factors carry each generator to its correspondent maximum power point. In these conditions, the total available power is PT = 1636 W, which corresponds to an output current with a maximum value of 10.5 A. During the first operation stage the three output voltage of each h-bridge module have the same amplitude. However, after the irradiance changes, the three voltages become different. This is only due to the ch&ge in the proportional factors, because, as it can be observed in Fig 8, with the change of irradiation the MPP DC voltage barely changes. This causes an output voltage with a higher distortion, as it is clearly reflected in a higher output current ripple.

Then, the proportionality factors of the voltages can be calculated by means of the correspondentpowers:

By means of using a linear scalar PWM modulation, the duty cycle of any k-th module is calculated as follows:

(9) In those applications where all the modules deliver the same power, the proportional factors have the same value equal to lln. Therefore, the duty cycles are: dk

.-%d ( k = l ...n )

-

‘DCk

*

However, in this application each module k-th has to deliver its own available power Pk. Therefore, the duty cycle calculation is:

Fig. 7 shows a block diagram of the modulator corresponding to the k-th module.

RESULTS VI. SIMULATION

vHT

x

The main features of the simulated power circuit are the following ones: - Number of modules in series: 3 - Coupling inductance: 0,5 mH - Grid characteristics: 220V I 5 0 Hz - DC link capacitors: 2mF - Switching frequency: 2000 Hz Each one of the photovoltaic arrays is a series connection of 8 panels of the model BP-585F (BP Solar). The operation features of each generator under standard conditions (1000 W/m2 irradiance and 25 “C ambient temperature) are the following ones: - Total open circuit voltage: 178 V - Short-circuit current: 5A - MPP Voltage: 144 V - MPP Current: 4,72 A - MPP Power: 680 W A simulation experiment that shows the behaviour of the system under irradiance steps has been carried out. From the beginning of the experiment to the moment F2s, the irradiance of the three solar arrays is the same with a value of 1000 W/m2. At that moment, the irradiance changes in the second array to 800 W/m2 and in the third one to 600 Wlm2. In these new work conditions, the maximum available power and their correspondents V-I points are shown in Fig. 8. This stage continues until t=3s, where the system resumes to initial conditions. Fig. 9 shows the results of the output voltage and current, power of each module, duty cycles and proportional

734

..., s 4 k

Switching

Fig. 7. PWM modulator for k-th module 200

,

0

1



*.

I

2

‘t.

3



4

1

5

PV current (A)

Fig. 8. V-I characteristic under different irradiance conditions

,

””“W

-600,O

I



I

15,O

10.0 5.0

0.0 -5.0 -10.0

700,O 650 0 ....................................

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

600 0 ................................... 550 0 ..................................

........................................................................................................................... ....................................... ,....... ............................................................................

500 0 ...................................

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

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

450 0

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

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

4oo 0

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

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

350 0 .................................. 300.0

I

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

,...................................................................................

,

,

0,50 0 45 ....................................

I

:I

,.....................................................................................

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

0 40 ....................................

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

0 35 ...................................

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

0 30 ...................................

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

0 25 ...................................

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

0.15

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

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

1

0.20

I 0.25

0,30

0,35

0.40

0,45

0.50

0,55

0,60

0,65

0,70

Time (9)

Fig. 9. Simulation results of an irradiance step test experiment

VI. CONCLUSIONS It is demonstrated an improvement in the global efficiency of the photovoltaic generator when several arrays with lower amount of panels are used instead of a unique array with the total amount of panels. Moreover, each array will have its specific maximum point power. To take advantage of all the available power, it is necessary to use modular conversion structures with independent MPPT controls. In this paper a new control technique for the cascaded H-Bridge multilevel converter is introduced. Its main feature is to allow each module to deliver its maximum available power calculated for an independent MPPT control. Moreover, due to its multilevel nature, the output voltages are synthesised by means of several voltage levels, whose number increases when the number of single-phase H-bridge modules series

connected increases. This is an important advantage with respect to others structures due to the reduction of common mode perturbations, harmonic voltage and current amplitudes, etc.

REFERENCES U1 [2]

PI [4l

735

G.R. Walker, P.C. Semia, “Cascaded DC-DC Converter Connection of Photovoltaic Modules”. PESC 02 Conference Proceedings. M. Calais, V.G. Agelidis, M. Meinhardt, “Multilevel Converters For Single-phase Grid Connected Photovoltaic Systems: An Overview”. Solar Energy. Vol66, No5, pp 325-335. 1999. H. Valderrama-Blavi, L. Martinez-Salamero, C. Alonso, J. Maix6Altks, “Nuevas estructuras de Onduladores Fotovoltaicos”, SAAEI 0 1 proceedings. 2001 D. B. Snyman and J. H. R. Enslin, “An experimental evaluation of MPPT converter topologies for PV installations”, Renewable Energy, vol. 3, no. 8, pp. 841-848, 1993.

Suggest Documents