Available online at www.sciencedirect.

com

Energy Procedia 34 (2013) 351 – 361

10th Eco-Energy and Materials Science and Engineering

(EMSES2012)

Control of Parallel-connected AC to DC Converter with

Droop Technique for DC Microgrid Application
C. Tarasantisuka*, V. Chunkaga and P. Thounthongb
a
Department of Electrical Engineering, Faculty of Engineering,

Thailand E-mail: [email protected]

b
Department of Teacher Training in Electrical Engineering, Faculty of Technical Education

Thailand.

Abstract

This paper presents the control of a single-phase ac to dc converter using the rotating reference frame
control methodology wherein the phase angle obtained from a single-phase phase-locked loop is used
for axis-transformation. The I d , I q currents and output voltage of the converter are then controlled by PI
control strategy. The dc microgrid under study consists of two modules of the ac to dc converters
connected in parallel topology using droop control technique. The simulation results based on
Matlab/Simulink illustrate that I d , I q currents can be regulated by the designed PI controllers. Therefore,
zero steady state error and simple power calculation are achieved. Also, real power sharing between the
two converters is fulfilled by the parallel-connected converter topology with droop controller in such a dc
microgrid system, while the output voltage is maintained at 600 V by a PI controller at the secondary
control level.

©2013
© 2013TheThe Authors.
Authors. Published
Published by Elsevier
by Elsevier B.V.
B.V. Open access under CC BY-NC-ND license.
Selection
Selectionand peer-review
and/or under responsibility
peer-review of COE of Sustainalble
under responsibility of COE of Energy System,Energy
Sustainable Rajamangala University
System, of Technology
Rajamangala
Thanyaburi
University(RMUTT)
of Technology Thanyaburi(RMUTT)
Keywords: Phase Locked Loop (PLL); Single-phase ac to dc converter; Parallel-connected converter and Droop technique

* Corresponding author. Tel.: +668-9922-6826; fax: +662-287-9629

E-mail address: [email protected]

1876-6102 © 2013 The Authors. Published by Elsevier B.V. Open access under CC BY-NC-ND license.
Selection and peer-review under responsibility of COE of Sustainalble Energy System, Rajamangala University of Technology
Thanyaburi (RMUTT)
doi:10.1016/j.egypro.2013.06.763
352 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361

1. Introduction

Recently, renewable energy such as solar cell, fuel-cell and wind turbine plays an important role for
small distributed generator (DG) or microgrid (MG). The microgrid can be divided in two categories,
which is ac microgrid and dc microgrid. The ac microgrid is connected in ac grid and renewable energy
source with grid connected inverter. The dc microgrid is connected energy storage devices and renewable
energy source with dc to dc converter. In microgrid system, it is necessary to implement a hierarchical
control system. The hierarchical control of microgrid can be classified in three control levels:
1.1 Primary Control is devoted to control local variables, such as frequency and voltage. In addition
to other low-level controls, these local controllers deal with implementing and virtual impedance control
techniques.
1.2 Secondary Control works as a centralized automatic generation controller and compensate the
steady-state errors in the microgrid voltage and frequency, restoring their values to the nominal ones. The
dynamic control of secondary control is lower than primary control.
1.3 Tertiary control is responsible for optimization of microgrid operation and setting its interaction
with the distribution network by controlling the active and reactive power references for each DG unit.

To achieve high performance operation in terms of power generation and to improve the reliability of
bidirectional parallel converter system, droop control scheme is adopted in this paper. Using droop
technique, it is advantageous for hardware implementation and the communication between converter
modules is not required, when comparing to other existing methods, [1]. This paper is organized as
follows. Section 2 is devoted to a formal description of the dynamic model and control of the converter
consisting of single-phase ac to dc converter with phase locked loop (PLL), current and voltage control.
In Section 3, the current and voltage control loop designs by means of the so-called PI control strategy are
described. Section 4 presents a set of simulation results that demonstrates the viability of the proposed
design. Finally, concluding remarks are presented.

2. Converter Modeling and Control

2.1. The control of a single-phase ac to dc converter

 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361 353

AC TO DC CONVERTER

SA1 SB1

is L
Cdc Z dc
Vdc

vs

SA2 SB2

Modulator
PLL L, I q , Vm L, I d

I d*
, PI PI
,
Vdcref
d, q I q* 0
Ts / 4
e d, q PI

Fig. 1. The control structure of a single-phase full-bridge ac to dc converter

The control system architecture of the full-bridge single-phase converter is depicted in Fig. 1. Such a
system can be applied for a dc microgrid as it can accomplish bidirectional power flow. Moreover, this
allows the use of a PLL for synchronization to the utility grid as well as for phase angle calculation to be
used for axis-transformation. From Fig. 1, it can be also seen that the reference signal for implementing
the single-phase sine pulse width modulation (SPWM) is then set according to the obtained phase angle.
As seen, the control structure is configured as an outer constant output voltage (Vdc ) loop regulated by a
PI controller that determines the reference value of the currents for an inner current loop as follows, [5]

I d* I * and I q* 0 (1)
i* I sin( t ) and i* is* I cos( t ) (2)

The determined I d* , I q* are the dc currents, thus the dq axis-transformation is required for the
measured input current is . The orthogonal imaginary in common approaches are obtained by T / 4
delay for i , i currents, such that the measured signal is in the desired quantities, while the phase angle is
derived from the PLL. The resulting I d , I q currents are then compared to I d* , I q* for computing a control
signal by the PI controller. Controlling such a type 0 system through the PI control, zero steady-state error
can be achieved. However, using this transformation in a single-phase power system, the correction term
of decoupling the axes is needed, [5], hence the dq transformation can be further used for
constructing a SPWM signal.
354 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361

Vpwmd LI q V and Vpwmq LI d . (3)

Before the controllers are designed, the converter system needs to be modeled. Here, the switching
model of the full-bridge converter where the dynamics of the module is ,Vdc are taken into account can be
described as, [6]

dis Rs 1
is vs Vdc (4)
dt L L
dVdc 1 1
io is (5)
dt Cdc Cdc

where is ,Vdc are the dynamic variables of the converter and { 1,1} represents switch state of all
- Note that in this system both

2.2. Single-phase PLL

The PLL control system taken into account in this paper is shown in Fig. 2 and its closed-loop transfer
function can be obtained as
Kp
Kps
( s) LF ( s) Ti (6)
H ( s)
( s) 1 LF ( s) 2
Kp
s Kps
Ti
From (6), one can observe that the transfer function is a typical second-order low-pass filter that can
be used for disturbance attenuation and noise rejection. Comparing to the standard second-order
dynamical system, we get

Kp K pTi
n and (7)
Ti 2

From its step response, we can estimate the settling time t s within 2% of the final value as

ts 4.6 (8)
1
Where . From (6) and (8), the desired t s can be satisfied through the use of symmetrical
n

optimum 0.7 technique for the controller design, by setting K p and Ti as

9.2 2 ts 2
Kp 2 n and Ti (9)
ts n 2.3
 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361 355

LF c VCO
vs v vq vf
1 1
kp 1
Ti s s

Delay v dq vd
T /4
vs

Fig. 2. A single-phase PLL control system block diagram

2.3. Parallel-connected converters using droop control

After designing the current and voltage loops, droop controller in primary control level should be
employed. In parallel ac to dc converter system, droop effect is shown by means of dc current I dc
sharing. While we calculate the virtual resistor ( Rd ) from an acceptable range of dc voltage as

Rd I dc ( FL) (Vdc max Vdc min ) (10)

Rd (Vdc max Vdc min ) / I dc ( FL ) (11)

where Vdc max and Vdc min denote respectively maximum and minimum dc voltage, and I dc ( FL )
maximum full load current. PI control is adopted in dc voltage secondary control, as shown in (12)

Kiv 2
GPI K PV 2 (12)
s

The virtual resistor is selected to guarantee a maximum of 3% DC voltage deviation of it normal

value, substitute Vdc max , Vdc min and I dc ( FL ) in (11), thus the Rd is 4 .
356 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361

Low Bandwidth vdcref

Comunication * InerLoops vdc
vo v dc M
Voltage Loop Current Loop PWM Generator Ac to dc
I
and Driver converter1
*
vMG C
GMG ( s ) id c R
Virtual resistance O
vMG G
RD
Secondary Control R
Primary Control (Droop Method)
I
D
vdcref
* InerLoops vdc
vo v dc
Voltage Loop PWM Generator Ac to dc
Current Loop
and Driver converter 2
id c
Virtual resistance
RD
Primary Control (Droop Method)

Fig. 3. Parallel-connected converters using droop technique

3. Control system design

3.1. Current and constant voltage control loop designs

From Fig. 2, the switching model of the full-bridge ac to dc converter can be expressed as, [2], [8]

vac sab vdc , idc sab iL (13)

By applying a linearized small-signal model around some operating point to (13), we get an average
model [2], [8],

vac dab vac , idc dab iL (14)

From the small-signal analysis of (14), we get

vac vac (dab dab )(vac vac ) (15)

idc idc (dab dab )( iL iL ) (16)

Specifically, when taking into account only the small-signal perturbation, we get vac , idc , vdc as
follows,

vac dab vac dab vdc , idc dabiL dab iL (17)

Z dc Z dc
vac iL sL , vdc idc , vdc idc (18)
sZ dc Cdc 1 sZ dc Cdc 1
 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361 357

Also, the voltage loop of the converter system can be described as

Z dc
d ab (d ab iL d abiL ) d ab vdc iL sL 0 (19)
sZ dc Cdc 1

Ignoring the 100 Hz output voltage ripple at the double frequency of the input frequency,

vdc Vdc vripple Vdc , dab Z dc iL Vc vripple Vdc (20)

dab Zdc iL Vc vripple Vdc (21)
iL 2 sZ dc Cdc
Gid Vdc (22)
d ab d Z dc Ls LZ dc Cdc s 2
2
ab

The current loop transfer function Gid can be simplified as

iL 1
Gid Vdc (23)
d ab sL

Here, the transfer function of the time delay is randomly selected as

1
GDI ( s) (24)
1 1.5Ts

Also, the balanced power of the converter can be written as

Vacrms Vacrms
Vacrms I acrms Vdc I dc , I acrms Vc , I acrms Vc (25)
K K

Now, using the small-signal perturbation analysis to (26)

2
Vacrms
(Vc vc ) (Vdc vdc )( I c idc ) , (26)
K
we get
2M VacRMS 1
idc vac vc vdc , (27)
ro KM ro
Vdc Vdc
where ro ,M , and Z dc RL (28)
I dc Vacrms
Therefore, we obtain the voltage loop transfer function to be used for PI regulator design as

udc VacRMS RL
Gvdc . (29)
vc KM 2 sCdc RL

From (22), Bode plots of Gid at different values of duty cycle {0.2,0.5, 0.7,1} are shown in Fig. 4(a).
Compared to (23), it is easy to see that the simplified model of Gid gives the magnitude plot at high
358 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361

frequency exactly the same as in (22). As a result, we therefore use the reduced model for the PI
compensator design of the current control loop based on the frequency domain method. Bode plots of the
compensated system are shown in Fig. 4(b). Figures 5(a)-(b) illustrate the step responses of the current
and voltage, respectively.

Eq. 23

Eq. 22

(a) (b)
Fig. 4. (a) Frequency response of the uncompensated system
(b) Frequency response of the compensated system

(a) (b)

Fig. 5. (a) Step response of the current loop

(b) Step response of the voltage loop
 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361 359

4. Simulation results

Table 1. Parallel connected single-phase Ac to dc converter parameters

Parameters Value Unit

Input voltage (vs ) , output voltage (Vdc ) 220,600 V
Input inductance ( L) 1.8 mH
Load ( RL ) 160
Output capacitance (Cdc ) 1,100 F
Switching frequency ( f s ) 10 kHz
PI controller of PLL ( K p , Ki ) 0.07, 0.15
PI controller of current loop ( K pc , Kic ) 0.25,0.1
PI controller, voltage loop ( K pv , Kiv ) 0.04, 0.5
Droop coefficient Resistance ( Rd ) 4
PI controller of secondary control loop ( K pv 2 , Kiv 2 ) 0.0032,8
Sampling frequency 200 kHz
Solver type Fixed -step

Consider a single-phase parallel-connected ac to dc converter with two modules by the system

parameter values listed in Table 1. The system based on the primary current and voltage control loops to
be simulated is implemented through the use of the switching model in Matlab/Simulink environment.
Figure 6(a) shows the simulation results of the PLL. It can be seen that the determined phase angle is
employed in axis-transformation. Fig. 6(b)-(d) show the simulation results of the parallel converters. The
input current and voltage evolutions of the circuits are provided in Fig. 6(b). Clearly, such current and
voltage are in phase, thus power factor is nearly unity. The id , iq from the first converter module are also
provided in Fig. 6(c). The currents have large overshoots during initial state due to the effect of time
delay T / 4 component as in fig. 6 (d). Also, one can observe that at 1.6 second the secondary control can
effectively restore back the output voltage to 600 V as shown in Fig. 7.

65
frequency (Hz)

60

55

50

45
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

6
Theta

0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

400

200
Voltage (V)

-200

-400
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

(a) (b)
360 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361

Delay effect

(c) (d)

Fig. 6. Simulation results of the parallel converter with two modules using droop technique
(a) PLL simulation results
(b) Input current and voltage (is ( A), vs (V ) /10)
(c) I q ( A), I d ( A) currents
(d) i ( A), i ( A) currents

Fig. 7. An output voltage Vdc (V ) when the secondary control is active at 1.6 second
 C. Tarasantisuk et al. / Energy Procedia 34 (2013) 351 – 361 361

5. Conclusion

This paper has presented a droop technique that facilitates the control problem encountered in a
parallel-connected ac to dc converter with two modules. Within the proposed control design, the rotating
reference frame control has successfully used to regulate the currents in rotating axes (d, q) through the
use of a PI controller. The feasibility of the proposed design and its performance has been validated using
a Matlab/Simulink-based simulation. The results show that the parallel converter with droop control
exhibits good performance in real power sharing as well as maintaining the output voltage level at 600 V
while remaining the quality of input current. In further work that is underway, the results presented here
will be extended to address the fundamental challenge of its application as a power transfer unit between
ac and dc microgrids, i.e., hybrid microgrid.

Acknowledgements

This work has been supported by Ph.D. student scholarship of Coordinating Center for Thai
Government Science and Technology Scholarship Students (CSTS)

References

[1] G -controlled AC and DC microgrids a general

approach toward standardization , Vol. 58, No. 1, pp. 158-172, 2011.
[2] Dong D., Thacker T., Burgos R., B
multi-function single- EPE '09. , pp.1-10, 2009.
[3] IEEE Transactions
on, Smart Grid, October 2009. Vol. 2, Issue 2, pp. 278 286.
[4]
Distributed P IEEE Trans. on Ind. Electronics, Vol. 53, Oct. 2006 pp.1398 1409.
[5] -Phase
PESC 2005, pp.1377 1381.
[6] Pires
IEEE Transactions on Education, Vol. 45, Issue: 3, pp. 253 261.
[7] Xiaonan Lu, Guerrero J. M., Teodorescu R., Kanares T., K. Sun and L. H -connected Bidirectional
AC- ECCE 2011, pp. 4153 4160.
[8] - Virginia
Polytechnic Institute and State University Thesis