Dual-Mode Controller for MPPT in Single-Stage

Grid-Connected Photovoltaic Inverters

Gamal M. Dousoky1,2), Masahito Shoyama3), and Haitham Abu‐Rub1)
1) Texas A&M University at QATAR
2) Minia University, EGYPT
3) Kyushu University, JAPAN

Abstract—This paper presents a dual-mode controller to

improve the performance of maximum power point tracking in
grid-connected photovoltaic inverters. Most of grid-connected PV
applications require a controllable active/reactive power
generation. The proposed method employs both of the load angle
and the inverter modulation index into two control modes: Active
power mode, and reactive power mode. Such technique enables
an efficient and flexible active/reactive power control beside
MPPT function using a single-stage VSI. A simulation model is
developed for the overall system to investigate the performance Fig. 1. A generator connected to the power grid.
of the proposed dual-mode controller. Furthermore, the
controller is designed and implemented using a digital signal
processor (DSP). Algorithms are configured on a fixed point DSP inverter (VSI). The paper is organized as follows: the main idea
TMS320F2812. Moreover, the controller is simulated in of the proposed dual-mode controller is explained in the
MATLAB/Simulink environment and is experimentally following section. Then, the simulation and digital
investigated using a breadboard built especially for the case implementations of the technique are described. After that, the
study. The results show good dynamic and steady-state details of the experimental test circuit, results and discussions
performances. are presented. Finally, conclusions are addressed.
Keywords—Renewable energy systems; Photovoltaic; Power
conditioning; Voltage Source Inverters; Energy Efficiency; DSP. II. PROPOSED DUAL-MODE CONTROLLER
I. INTRODUCTION Figure 1 shows an equivalent diagram of a grid-connected
generation system, and its simplified vector diagram. The
A cost-effective deployment of grid-connected PV systems generator output voltage is equal to the grid voltage and the
can be achieved by minimizing the initial investment cost generator is the source of active power (Pge) and reactive power
required to purchase and install the PV system components (Qge). In order to control both of Pge and Qge, either/both of the
(e.g., PV modules, dc/ac inverters, etc.), maximizing the load angle (δ) or/and the modulation index (mi) can be used
amount of energy injected into the electric grid, and increasing according to (1) and (2):
its reliability [1]‒[5].
Grid-connected PV systems use dc/ac inverters to inject the Pge = Ege Vgr sin (δ) / Xg (1)
PV power into the utility grid and control both of active and 2
Qge = [(Vgr) –Ege Vgr cos (δ)] / Xg (2)
reactive powers [6]. It is crucial to operate the PV energy
conversion systems near the maximum power point (MPP) to where;
increase the efficiency of the PV system [7].
Voltages are expressed in rms values.
Many MPP tracking (MPPT) methods have been developed mi is the inverter modulation index.
and implemented. The methods vary in complexity, required δ is the load angle; the phase angle of the sinusoidal
sensors, convergence speed, cost, range of effectiveness, pulse-width modulation (SPWM) control signal
implementation hardware, popularity, and in other respects [8]. (w.r.t. the power grid voltage phase angle).
The general requirements for MPPT are simplicity and low
cost, quick tracking under changing conditions, and small
output power fluctuation. Efficient methods to solve this The above principle is usually used in case of conventional
problem become crucially important [7]‒[8]. The need to power generation systems [12]. Moreover, the authors applied
improve the efficiency and reduce the size and the cost of both this principle to maximize the power produced by a grid-
the inverter and the output filter encouraged more research into connected PV system; as presented in [13]. Their investigations
using different inverter topologies [10]‒[11]. concluded that tracking with δ exhibits steady state oscillations
around the MPP. Moreover, tracking with mi has a stable
This paper introduces a novel dual-mode controller that steady state performance. On the other hand, the latter has a
enables an efficient and flexible active/reactive power control poor dynamic performance.
beside an MPPT function using a single-stage voltage source

978-1-4799-0224-8/13/$31.00 ©2013 IEEE 665

Therefore, a dual-mode controller is introduced in this  Reactive power mode: The controller uses mi to track
paper. It employs both of the control variables (δ, and mi) to MPP, while holding δ constant. The inverter
attain a good dynamic and steady state performances. modulation index directly controls the magnitude of
Moreover, using both of δ and mi enables an efficient and the inverter ac output voltage (Ege); which in turn
flexible control of both Pge and Qge. Fig. 2 shows the proposed affects Qge deeply in this mode. This mode is applied at
technique’s drive synthesis flowchart. The MPPT controller the steady-state operation.
uses the incremental-conductance (INC) method [14]-[15] for
tracking the MPP. It alters between two main modes of The mechanism of switching the MPPT control mode
operation: between the active power mode and the reactive power mode is
based‒in this investigation‒on a certain boundary value of the
 Active power mode: The controller uses δ to track load angle (δB) for the sake of testing the proposed dual-mode
MPP, while mi constant. The load angle usually has a MPPT. In other words, when the load angle increases more
small value. In such band of angles, the sine function than a threshold value (δB), the MPPT changes the control
has a sharp slope. Therefore, Pge is deeply affected in mode from active to reactive power mode, as illustrated in Fig.
this mode. Thus, this mode is applied at the start-up 2. Furthermore, the decision to change the control mode may
period.

Start
mi(n-1) = mi(n)
δ(n-1) = δ(n)

Read ipv, vpv

IPV(n-1) = IPV(n), VPV(n-1) = VPV(n)

IPV(n) = ipv, VPV(n) = vpv
dIPV = IPV(n) - IPV(n-1), dVPV = VPV(n) - VPV(n-1)

if
Yes dI PV I PV No
 ≥0
dVPV VPV

S = -1 S = +1

Yes change No
Reactive Active
mode?
power mode power mode
δ (n) = δ(n-1) mi(n) = mi(n-1)
mi(n) = mi(n-1) + S ×Δmi δ (n) = δ(n-1) + S ×Δδ

output mi(n), δ(n)

S: Sign of the change in the control variable {+1, -1}

Fig. 2. The proposed dual-mode MPPT’s drive synthesis flowchart.

666
be taken in different ways‒depending on the controller features active to reactive to use the PV system as reactive
and target including, but not limited to, the following power compensator.
methodologies:
 Depending on the rate of change of the PV panels
 Using a P-Q controller: When achieving the active output power (ΔPpv): When ΔPpv decreases below a
power target, it takes the decision to change the mode certain limit, it takes the decision to change the mode
from active to reactive power mode. from active to reactive power mode.
 Depending on the solar radiation conditions: When  Depending on the PV panels output voltage (Vpv): When
the solar radiation decreases below a certain limit, it Vpv decreases below a certain limit, it takes the decision
takes the decision to change the power mode from to change the mode from active to reactive power mode.

VPV, IPV, PPV

δ, mi
V δ

I mi
Three Phase Phase Angle
SPWM Control Circuit
INC MPPT δ
ctrl A A_lead
A_out mi
A
A_in
ctrl B B_lead
B B_out B_in
3ph SPWM
Discrete,
Ts = 1e-006 s. ctrl C C_lead
C
C_in

C_out
N_in
powergui Carrier

v
Three-Phase
g
+ - Voltage Sensor
+
Triangular wave
A
generator
H H + + -i L1 L2
Cin A A A A A

T B B B B B B N

298.15 T - - C C C C C

C Power Grid
300W PV Panel
Universal Bridge
Cf

Fig. 3. Simulation model using SIMULINK tool boxes.

1 1
1
delta
z

I 1
z
1

V
1
z

calculation of 0.01 2

dI PV I Mode
 PV
m_i

dV PV VPV selector
0
1
z

Fig. 4. Implementation model of the proposed tracker’s algorithm (flowchart is shown in Fig. 2).

667
III. SIMULATION AND DIGITAL IMPLEMENTATION IV. EXPERIMENTAL INVESTIGATION AND RESULTS
A simulation model is developed for the overall system to The proposed dual-mode controller is experimentally
investigate the performance of the proposed dual-mode investigated. A PV power simulator Agilent E4360 is
controller using the INC MPPT method (see Fig. 3). It consists programmed to simulate the characteristic of the PV panel. The
of PV model, dc-ac six switches inverter, power grid, LCL experimental setup for the proposed dual-mode single-stage
filter, and a controller. The controller employs an INC MPPT, grid-connected PV system is shown in Fig. 6, and the overall
a three-phase voltage sensor, a phase angle control circuit, a system parameters addressed in Table I. The overall system
triangular wave generator, and a sinusoidal pulse-width consists mainly of the following units:
modulator (SPWM). The INC MPPT is designed to track the
MPPT using any of the previously-addressed tracking modes. 1. An Agilent E4360 PV simulator; which is programmed
to simulate the characteristic of a 300W PV panel.
The proposed dual-mode controller is designed and 2. PV voltage and current sensors
implemented using a digital signal processor (DSP).
Algorithms are configured on fixed point DSP TMS320F2812 3. A digital controller (fixed-point DSP TMS320F2812);
using MATLAB Simulink toolboxes. The implementation is which contains 12-bit A/D converters.
developed in MATLAB Simulink environment and then the 4. Isolation and driver circuit.
target is programmed using code composer 3.1. Fig. 4 shows 5. Three-phase six-switch inverter.
an implementation model of the proposed tracker’s algorithm. 6. LCL filter.
The tracking frequency is 6 Hz that means the tracker 7. Three-phase voltage sensors to sense the grid voltages.
updates the control variables every 10 AC cycles. Both 8. Synchronizing unit/switch to enable connecting with
schemes are tested under the same conditions. A fixed step size the power grid.
perturbations for mi, and for δ are taken as (Δmi = 0.01, Δδ =
1◦). Table I illustrates the overall system parameters. Fig. 5 A Lecroy 424 WaveSurfer oscilloscope is used to record
shows the simulated PV voltage, current, power, δ, mi and the experimental results. A 1/50 differential probe is used to
during MPP tracking period for the dual-mode controller. It is measure the voltage. The tracking waveforms of the PV output
clear that the proposed dual-mode controller can successfully voltage, current, and power using the proposed dual-mode
track the maximum available power of the PV (300W). MPPT at different tracking techniques/modes are presented in
Fig. 7.

TABLE I. OVERALL SYSTEM PARAMETERS

PV Voltage (V)

100
Symbol Description Value
50 VOC Open circuit voltage (V) 120
0
Vmpp Voltage at max power (V) 100
0 1 2 3 4 5 6 7 8 9 10 ISC Short circuit current (A) 3. 5
Time (S)
Impp Current at max power (A) 3.0
3 Pmpp Maximum power (W) 300
PV Current (A)

2 Cin Input capacitor (uF) 2000

1 L1, L2 Filter inductors (mH) 5.3, 1.65
0 Cf Filter capacitor (uF) 4
0 1 2 3 4 5 6 7 8 9 10
Time (S)
Fs Switching frequency (kHz) 21.6
Vgr Grid line voltage (Vrms) 200
PV Power (W)

300

200
100

0
0 1 2 3 4 5 6 7 8 9 10
Time (S)
Phase Angle (delta)

0 1 2 3 4 5 6 7 8 9 10
Time (S)
Modilation Index
Modulation (mi)
Index (mi)

0 1 2 3 4 5 6 7 8 9 10
Time (S)

Fig. 5. Simulation results: starting waveforms using the proposed dual-mode

MPPT technique. Fig. 6. Experimental setup for the proposed dual-mode single-stage grid-
connected PV system.

668
VPV [25V/div] VPV [25V/div]

IPV [1A/div] IPV [1A/div]

Active
Active power mode only power mode Reactive power mode

PPV [50W/div]
PPV [50W/div]

(a) Tracking with δ at constant mi (Basic: Good start-up performance, (d) Tracking with δ up to δm=3◦, then tracking with mi (Best: Efficient
however, poor steady-state performance). start-up and steady state performances).

VPV [25V/div] VPV [25V/div]

IPV [1A/div] IPV [1A/div]

Active
Reactive power mode only power mode Reactive power mode

PPV [50W/div] PPV [50W/div]

(b) Tracking with mi at constant δ (Basic: Stable steady-state (e) Tracking with δ up to δm=5◦, then tracking with mi (Bad: Any small
performance, however, poor start-up performance). change in mi causes large perturbations in Vpv and Ppv).

VPV [25V/div]
VPV [25V/div]

IPV [1A/div]
IPV [1A/div]
Active
power mode
Reactive power mode
Both Active and Reactive power modes are working

Overlap
PPV [50W/div] PPV [50W/div]

(f) Overlap tracking: Use δ up to δm=3◦, and use mi from δm=2◦ (Good:
(c) Tracking with both of δ and mi instantaneously (Bad: The big change accelerates reaching the MPP during the start-up period without
in the independent variables (δ and mi) causes large perturbations in causing perturbations in Vpv or in Ppv at steady-state).
Vpv and Ppv).

Fig. 7. The experimental tracking waveforms using the proposed dual-mode MPPT, compared to other different tracking techniques/modes (a 1/50 differential
probe was used to measure the voltage).

669
As shown in Fig. 7(a), the basic tracking with δ at constant [3] G. M. Dousoky, A. M. El-Sayed, and M. Shoyama, " Increasing energy-
mi achieves a good start-up performance and, contrary, a poor efficiency in solar radiation trackers for Photovoltaic arrays", In Proc.
IEEE-ECCE 2012, Sept. 2012, pp. 4113-4120.
steady-state performance. Furthermore, in Fig. 7(b), the basic
[4] J. M. Fife, M. Scharf, S. G. Hummel, and R. W. Morris, “Field
tracking with mi at constant δ achieves a stable steady-state reliability analysis methods for photovoltaic inverters,” in Proc. 35th
performance and, contrary, a poor start-up performance. IEEE Photovoltaic Spec. Conf. (PVSC), Jun. 2010, pp. 2767–2772
On the other hand, when tracking with both of δ and mi [5] E. Koutroulis and F. Blaabjerg, “Design Optimization of
Transformerless Grid-Connected PV Inverters Including Reliability,”
instantaneously the MPPT attains a bad performance (see Fig. IEEE Trans. on Power Electron, vol. 28, no. 1, pp. 325-335, Jan. 2013.
7(c)). This is due to the fact that the big change in the [6] S. Zengin, F. Deveci, and M. Boztepe, “Decoupling Capacitor Selection
independent variables (δ and mi) causes large perturbations in in DCM Flyback PV Microinverters Considering Harmonic Distortion,”
Vpv and in Ppv. The same occurs when tracking with mi at large IEEE Trans. on Power Electron, vol. 28, no. 2, pp. 816-825, Feb. 2013.
value of δ (see Fig. 7(e)). [7] A. M. Kassem, “Maximum power point tracker control design of a
photovoltaic-generator powered DC motor-pump system based on
The best performance is obtained when tracking with δ up artificial neural networks,” Australian Journal of Electrical &
to δm=3◦, then tracking with mi (shown in Fig. 7(d)), where the Electronics Engineering, Vol. 9, No. 2, pp. 127-136, 2012.
MPPT attains an efficient start-up and steady state [8] T. Esram and P. L. Chapman, “Comparison of Photovoltaic Array
performances. Moreover, overlap tracking achieves a good Maximum Power Point Tracking Techniques,” IEEE Trans. on Energy
performance (see Fig. 7(f): It accelerates reaching the MPP Conv., vol. 22, no. 2, pp. 439-449, Jun. 2007.
during the start-up period without causing perturbations in Vpv [9] G. M. Dousoky, E. M. Ahmed, and M. Shoyama, "Current-sensorless
MPPT with DC-DC boost converter for Photovoltaic battery chargers",
or in Ppv at steady-state). In Proc. IEEE-ECCE 2012, Sept. 2012, pp. 1607-1614.
[10] M. A. Abusara and S. M. Sharkh, “Design and Control of a Grid-
V. CONCLUSIONS Connected Interleaved Inverter,” IEEE Trans. on Power Electron., vol.
28, no. 2, pp. 748-764, Feb. 2013.
A novel dual-mode MPPT controller is proposed for single-
[11] G. M. Dousoky and M. Shoyama, " Current-sensorless power-angle-
stage grid-connected PV inverters. It employs both of the load based MPPT for single-stage grid-connected photovoltaic voltage-source
angle and the inverter modulation index into two control inverters", In Proc. IEEE-ECCE 2013, Sept. 2013, pp. 1-7, in press.
modes. The mechanism to switch between the control-modes is [12] R. Strzelecki and G. Benysek, (2008) Power Electronics in Smart
illustrated in detail. Both of the simulation and experimental Electrical Energy Networks. Chapter13: Integration of Distributed
results proved its efficient dynamic and steady-state Generation with Electrical Power System. Springer-Verlag London
performances and applicability to track the MPP in a grid- Limited, pp 303-325.
connected PV energy system. Such technique enables an [13] G. M. Dousoky, E. M. Ahmed, and M. Shoyama, “MPPT Schemes for
Single-Stage Three-Phase Grid-Connected Photovoltaic Voltage-Source
efficient and flexible active/reactive power control beside Inverters”, In Proc. IEEE-ICIT 2013, Feb. 2013, pp. 600–605.
MPPT function using a single-stage inverter. [14] K. H. Hussein, I. Muta, T. Hshino, and M. Osakada, “Maximum
photovoltaic power tracking: An Algorithm for Rapidly Changing
REFERENCES Atmospheric Conditions,” In Proc. IEE Generation, Transmission and
Distribution, 1995, vol. 142, pp. 59-64.
[1] O. Haillant, “Accelerated weathering testing principles to estimate the
[15] A. Safari, S. Mekhilef, “Simulation and Hardware Implementation of
service life of organic PV modules,” Solar Energy Mater. Solar Cells,
Incremental Conductance MPPT with Direct Control Method Using Cuk
vol. 95, no. 5, pp. 1284–1292, May 2011.
Converter,” IEEE Trans. on Ind. Electron., vol. 58, no. 4, pp.1154-1161,
[2] Gamal Dousoky, On Intelligent Power Electronic Interface for April 2011.
Renewable Energy Systems: Design, Cost, Reliability Investigations,
VDM Verlag Publishing House Ltd., chap. 2-3, Feb. 2010.

670

Powered by TCPDF (www.tcpdf.org)