Balanced Active and Reactive Control Applied to a Grid Connected Five Level Inverter

This paper presents a balanced active and reactive power control, using a Phase Locked Loop for synchronization, and applied to a grid connected Five Level Inverter. The energy source of the system can be a photovoltaic generator or a wind turbine. We size the passive elements of the system and explain the value of the system architecture using a Five Level Inverter when compared to a classical grid connected system. We also compare the balanced active and reactive power control to an unbalanced active and reactive power control. The simulation results obtained by using Matlab Simulink and Simpowersystems are presented and discussed in this paper. Keywords—Balanced control; grid connected system; multilevel inverter


I. INTRODUCTION
Maximizing the power transfer from the energy source to the electrical grid and optimizing the system dynamics both depend on the used static converter technology and its control algorithms [1].In grid-connected system, we must use a filter to connect the inverter to the utility grid in order to obtain a good attenuation of high frequency harmonics generated by the inverter [1], [2], [7].This is the case of [1]- [3], [15], [17] and [18].However, using a -filter requires a sophisticated control, due to the increase in filter order.Insofar as it is no longer possible to control the system using traditional controllers ( : Proportional Integral), we must use other types of control such as cascade controllers, sliding mode control, fuzzy logic control as in [11], [12] and [16] .
To solve this problem, we propose an architecture of gridconnected system using a Five-Level Inverter ( ) with an interleaved -filter and controlled by a balancing -control.The use of this innovative structure using a in this domain should allow a decrease in filter volume and a gain in system reliability, and in energy quality.
So the objectives of this work are to analyze the gridconnected system using an with an interleaved -filter and to show the value of this architecture compared to the classical grid-connected system.We are particularly interested in balancing control and sizing of this system, in a context of use in Africa where the cost, reliability and energy quality issues of the system should be taken into account.The considered grid is a low voltage grid ( ) with a frequency equal to fifty hertz ( ) but likely to vary, in which a maximum power of five point two kilowatts (5.2 kW) is injected.Those are the standards of the national electricity company of Chad ( : Société Nationale d'Electricité du Tchad).This work is organized as follows: Section describes the system under study, Section presents the principle, Section presents the sizing of passive elements, and Section details the system's modeling and control.Finally, the simulation results and discussion are provided in Section .

II. DESCRIPTION OF THE SYSTEM
Fig. 1 shows the proposed architecture of the grid connected system and its control.
This system is composed of two parts: a continuous part ( ) and an alternative part ( ).The part is the energy source which can be a photovoltaic generator or a wind turbine.The part is composed by the , the utility grid and is controlled by a balanced active and reactive power control ( -control).A Phase-Locked Loop ( ) is used to synchronize the system with the grid.The balanced control is composed of two control loops : an external control loop to control the output current in the Park frame and an internal control loop for local balancing of the cell currents which is done in the frame.www.ijacsa.thesai.org

III. PRINCIPLE OF THE FIVE-LEVEL INVERTER ( )
The simplified diagram of the three-phase is shown in Fig. 2, where is the phase name of the .It is a parallel structure composed of four interleaved cells.As in a classical inverter, the two switches of the same inverter arm are controlled complementarily to avoid short circuits.
; Where { } is the cell number.
The different switching cells are shifted by and controlled with the same duty cycle [4], [5].Where and is the switching period.The three phases of the inverter ( ) are also shifted by .
To simplify, the study of the three-phase inverter can be reduced to a single phase.Fig. 3 shows the output voltage waveforms of the and classical inverter (two-level inverter) for one phase when a load is connected (For specifications of the + load, cf.Table 2).Contrary to the waveform of the classical two-level inverter (Fig. 3(b)), the waveform of the is more sinusoidal (Fig. 3(a)).Indeed, this waveform is less rich in harmonics.Moreover, these harmonics are at the apparent frequency which is equal to times the switching frequency, so they are easy to filter.We can therefore use small caliber passive elements to make the filter.Furthermore, we can see in Fig. 3(a) that the waveform of the voltage has five voltage levels (two alternating positive + two alternating negative + zero level); hence, the name of the Five Level Inverter ( ).

A. Grid modeling 1) Medium voltage model of the grid
The considered medium voltage line ( ) of the grid has a distance and a section www.ijacsa.thesai.org[8].So, the resistance of the medium voltage line ( is calculated in (1). (1) Where is the aluminum resistivity.In the case of Chad which is a hot country with an ambient temperature substantially equal to [8], we can consider a maximum temperature equal to .In addition, this temperature corresponds to the temperature indicated in the datasheet of the considered transformer [9].In this case, (if , ) and .
The linear reactance is equal to [8].So the grid medium voltage line inductance is given in ( 2). ( Where is the grid pulsation ( ).For , we obtain .

2) Low voltage model of the grid
To switch from the medium voltage model to the low voltage model, we use the transformer model seen Fig. 4.
On the other hand, Using ( 6), ( 4) becomes ( 7). ) of the transformer are presented in Fig. 4. The line capacitor is neglected here, so the grid impedance is only composed of its resistance and inductance.
To simplify, we set: , and .
The transformer line-to-phase voltages are: and [8].
Thus, using (8), we obtain the grid impedance parameters of the low voltage model: and .
So the module of the grid impedance is given in (9).
Where is the grid pulsation ( is the grid frequency).

For
, we obtain .

B. Impedance of the transformer
The model of the considered transformer is shown in Fig. 4. Some characteristics of this transformer are given in Table 1 [9].Where: is the magnetic resistance; is the magnetic reactance; is the active iron losses; is the reactive iron losses and is the magnetic inductance.
The active and reactive iron losses are calculated as follows: (13) (14) (15) On the other hand, characteristics of the Table 1 are given in the case of a three-phase system.For the active iron losses of one phase , we can consider the open circuit losses ( divided by three as shown in ( 16). ( So the reactive iron losses are calculated in (17). √ By injecting ( 16) in (10), (17) in (11) and using Table 1, we obtain: and .Finally, using (12) we obtain .

2) Short circuit test and serial elements calculation
The serial resistance and inductance of the transformer are calculated by the short circuit test.The formula of the serial resistance is given in (18).(18) Where is the rated secondary current. ( By ( 18) and ( 19) and using table I, we obtain .
The mainly parameters which allow to calculate the serial inductance of the transformer are the following: For the active short circuit losses of one phase , we can consider the losses due to load divided by three as shown in (22).Because the parameters of Table 1 are given in the case of a three-phase system.Where: is the secondary short circuit voltage; is the apparent short circuit losses; is the active short circuit losses; is the reactive short circuit losses and is the reactance of the transformer.By considering the above parameters, we obtain .
Note: We can neglect the magnetic resistance and inductance of the transformer because they have very high values and are connected in parallel (cf.Fig. 4).Thus the internal resistance and inductance of the transformer are equal to the serial resistance and the inductance: and .So the module of the transformer's impedance is given in (26).

C. Impedance of the Cable
We consider that the distance between the common connection point of the grid and our installation is equal to .For this short distance, the capacitor of the cable can be neglected and the model of the cable is only composed of a resistance and an inductance.In addition, a three-phase aluminum cable with a section equal to is selected.
The formula of the cable resistance is indicated in (27). ( Where is the aluminum resistivity ( at ).So we obtain .
And the formula of the cable inductance is given in (28). ( Where is the reactance of the cable. ( Where is the linear reactance of the cable ( ) [8].By injecting (29) in (28), we obtain .So the impedance of the cable is given in (30).

D. The Interleaved Inductance (Filter Inductance)
The four inductances of the interleaved -filter are identical.The value of these inductances is selected to obtain a www.ijacsa.thesai.orgcurrent harmonic distortion of less than five percent ( in order to respect the connection standard [10].In addition, nowadays there are photovoltaic inverters on the market with a lower than or equal to Therefore, we are focusing on those existing solutions.Fig. 5 shows the of current versus ( ) inductance.We can see in Fig. 5 that for , , so this inductance value will be selected.Therefore, we use four identical inductances to make the interleaved -filter ( ).
The passive elements of our system are summarized in Table 2. To compare filter volumes, the same inductances values are considered both in the classical inverter and in the .2, we calculate and compare the volumes of passive elements as follows [7], [13]:  In general: -Energy: (31)  Three-phase classical inverter ( :  Three-phase (q = 4): By ( 4) and ( 5), the volume ratio is follows: (36) Note: ⇒ , so the overall volume is reduced in the case of although many components are required.

A. Output Current Control (External Loop)
By considering the internal resistance of our system's components (filter, cable, transformer and grid), the simplified diagram of the grid-connected is shown in Fig. 6.Mathematical models of Fig. 6 are the following: -In the frame: Where , and are the inverter voltage, the grid voltage and the grid current respectively.
is the total inductance ( ) whereas is the total resistance ( ). www.ijacsa.thesai.org Our control system can be assimilated to a cascade control with an external loop to control the output current and an internal loop for the cell currents control.In the case of the two cascaded loops, the internal loop must be faster than the external loop [13].To calculate the parameters of the cell currents (internal loop), we have chosen a bandwidth smaller than ten times the switching frequency.We chose a value of for this frequency.Moreover, the damping factor chosen is which gives a fast response without overshooting [14].The total inductance and resistance values are and .For these parameters and with the defined bandwidth and damping factor values, we calculate the values of the parameters ( and ) by using (45) [14].So (45) becomes (46).

{ (46)
Were is the pulsation corresponding to the bandwidth; and are the parameters for the internal loop.We obtain and .

VI. SIMULATION RESULTS AND DISCUSSION
The simulation is done using Matlab Simulink and SimPower Systems by considering the specifications given in Table 2.
So in order to test the developed control, we set and is varied on time ( as follows: -for [0 1] s, -for [1 2] s, -and for [2 3] s, Where and are the reactive and active power references respectively.Some simulation results are shown in Fig. 8 to 14.We can see in Fig. 8 that the current waveform varies according to .This is normal because by varying we indirectly vary the current ( ) (This variation is also noticeable in Fig. 9 and 10).Thus, the developed control tracks the references and can be validated.Fig. 9 and 10 show the waveforms of the cell currents without and with balancing respectively.The balancing of cell currents is necessary to evenly distribute the output current between the different cells in order to increase the reliability of the system [4], [5], [14].Note that the output current is equal to times the cell current ( ).
Fig. 11 shows the waveform of the three-phase grid voltage where we can see that the voltage is sinusoidal and smooth.Therefore, our system injects current into the grid without disturbing it.
To check the condition of current frequency and grid frequency matching, we need a Phase Locked Loop ( ) to synchronize our system with the grid [6].Fig. 12 shows the superposition of the injected current into the grid of the phase , on the simple grid voltage curve of the same phase.We can see that the current ( ) and the grid voltage ( ) are in phase and have the same frequency ( ).
Fig. 13 shows the current and voltage Total Harmonic Distortion ( ).The of the current is equal to (cf.Fig. 13(a)) whereas the of the voltage is equal to (cf.Fig. 13(b)).Both respect the IEE1547 standard which requires to be smaller than for the first 50 harmonics [10].

VII. CONCLUSION
The sizing of passive elements and simulation results showed the value of the converter.The strengths of this inverter are the quality of the waveforms (see Fig. 3), the reliability of the system obtained by using several cells and the reduction of the output filter volume.The developed balanced -control allowed to follow the instructions while balancing the currents flowing through the switching cells of the .The balancing of cells currents is very important to increase the reliability of the system.The interleaved -filter has been sized to obtain a current equal to and a voltage equal to However, the disadvantage of the proposed architecture is the number of semiconductors used to make the .
Fortunately, this disadvantage can be mitigated by the reliability of the system and the low price of the semiconductors [7].
In the future, it would be interesting to take into account the impact of the neighbor's consumption on our system.

Fig. 4 .
Fig. 4. Simplified transformer model.By neglecting the losses in the transformer, we can write the power conservation relation in (3).
and low grid line-to-phase voltages respectively; and are the medium voltage side and low voltage side grid currents; and are the medium voltage side and low voltage side grid impedances ; and are the medium voltage side and low voltage side grid resistances; and are the medium voltage side and low voltage side grid inductances; And is the voltage ratio.Note: , , and are not represented in Fig. 4.Only the internal elements (

TABLE II .
SPECIFICATIONS OF THE PASSIVE ELEMENTS

Symbol Parameter Value Unit classical inverter FLI (q = 4)
Note: The maximum amplitude of output current shown in this table takes into account the current ripple.Its fundamental value is equal to :