4780 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO.

10, OCTOBER 2013

A Seamless Control Strategy of a Distributed

Generation Inverter for the Critical Load Safety
Under Strict Grid Disturbances
Tai-Sik Hwang, Student Member, IEEE, and Sung-Yeul Park, Member, IEEE

Abstract—This paper introduces a seamless grid interconnec-

tion control strategy for distributed generations (DG) with renew-
able energy. Based on the performance analysis of conventional
voltage and current control loops, sophisticated control strategies,
which can overcome the limited capability of the DG inverter con-
trol under mode transition conditions for unknown power plants,
are needed in order to protect critical loads and operate the DG
inverter without grid fault trip. The proposed control strategy con-
sists of a current controller and a feedforward voltage controller
to minimize the grid overvoltage. The feedforward voltage control
loop is added to the d–q axis current control loop. The proposed
control strategy reduces the overvoltage stress of the renewable
energy and the critical load under the grid fault. In addition, the
seamless operation of the DG inverter will also enhance the sta-
bility and reliability of the grid. Real-time digital simulator-based
hardware-in-the-loop experimental results and simulation results Fig. 1. DG inverter system configuration with operational modes.
show that the DG inverter achieves the seamless mode transition
under grid overvoltage conditions.
Index Terms—Critical loads, distributed generation (DG) in- In the literature, two categories of critical loads have been
verter, distributed generations, grid-connected (GC) mode, mode discussed. One is the grid-scale power loads requiring very high
transfer, mode transition, standalone (SA) mode. quality of power including medical equipment, semiconductor
industry, and broadcasting facilities [2]–[4]. The other is the aux-
iliary power system to supply ac power to the balance of power
I. INTRODUCTION
plant in renewable or energy storage power systems [5], [6]. A
HE use of the renewable energy is increasing rapidly at
T a growing rate. The growth in renewable generation is
expected to be 26% of the total generation growth from 2009
low-cost power electronics stage based to provide ride-through
capability for critical loads has been explored [2]. The use of a
micro high-temperature superconducting magnetic energy stor-
to 2035 in U.S. [1]. Therefore, utility companies have already age system has been proposed to support critical industrial loads
begun to take into account not only the conventional central- with a ride-through capability of around 20 cycles [3]. The mi-
ized power generation, transmission, and distribution, but also crowind energy conversion scheme with battery energy storage
renewable energy-based distributed generations (DGs). With was proposed as support for the critical load [4]. The dc–ac
increasing renewable DG, fast and stable mode transition tech- inverter in the current-controlled mode exchanges active and
nologies are substantial not only for sending power to the grid, reactive power, which provides continuous power to the criti-
but also for protecting DGs from grid fault conditions. Par- cal load under operational modes. A control algorithm for fault
ticularly, to supply power to the critical loads under any grid ride through with voltage compensation capability for the criti-
condition has become an important issue [2], because most crit- cal load is proposed in a three-phase utility-interactive inverter
ical loads are sensitive to voltage variations, which can make with a critical load [5]. The safeties of mechanical and electrical
the critical loads’ performance worse or shut down the system balance of power plants are very important and considered for
operation. critical loads in operating fuel cell power plant systems [6]–[8].
A new voltage sag compensator for powering critical loads in
electric distribution systems has been discussed in an ac–ac
Manuscript received July 20, 2012; revised October 30, 2012; accepted converter [9].
December 3, 2012. Date of current version March 15, 2013. This work was Fig. 1 shows a systematic configuration of the DG inverter
supported by the Northeast Utilities Foundation. Recommended for publication [10]. Usually, the DG has a dc–ac inverter-based power con-
by Associate Editor M. Ponce-Siva.
The authors are with the Department of Electrical and Computer Engi- ditioning system, which is used either to deliver ac power to
neering, University of Connecticut, Storrs, CT 06269-2157 USA (e-mail: the grid or to supply ac power to the load. In the grid-connected
tai-sik.hwang@engr.uconn.edu; supark@engr.uconn.edu). (GC) mode, DG inverters control the output current with respect
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. to the voltage phase angle to send ac power to the grid [10]. In
Digital Object Identifier 10.1109/TPEL.2012.2236578 the standalone (SA) mode, DG inverters supply ac power to
0885-8993/$31.00 © 2012 IEEE
HWANG AND PARK: SEAMLESS CONTROL STRATEGY OF A DISTRIBUTED GENERATION INVERTER FOR THE CRITICAL LOAD SAFETY
the critical or local load by regulating the output voltage. Two
physical switches connect or disconnect DGs. The mode switch
is turned ON and connected to the grid in the GC mode and
is turned OFF in the SA mode under grid fault conditions. A
static transfer switch or circuit breaker can be used as the mode
switch [11]. The point of common coupling (PCC) switch is
another protection switch. If the grid is under fault conditions
such as over/under voltage, over/under frequency, etc., then the
DG is disconnected from the grid for the protection of the crit-
ical load and the DG inverter. If the auxiliary power of the DG
inverter control board supplies from the grid as a critical load,
the DG inverter should operate in both GC and SA modes to pro-
vide uninterrupted and continuous power [12]. Therefore, it is
important that the DG inverter controller can detect exact fault
conditions and transfer the seamless operational mode within
allowable duration to reduce voltage and current spikes.
Seamless mode transition methods have been investigated
[13]–[24]. A seamless transfer algorithm can switch the in-
verter operation from the voltage control mode to the current
control mode and vice versa with minimum interruption to the
local load [19]. The mode switch helps in disconnecting the grid
within a half-line cycle. An indirect current control algorithm for
seamless transfer of utility-interactive voltage source inverters
has been proposed [20], [21]. With the proposed method, the DG
inverter can provide ac power in critical loads with a stable and
smooth output voltage during the whole transition time includ-
ing both clearing time and control mode change [20]. A seam-
less transfer of single-phase grid-interactive inverters between
GC and SA modes was presented [22]. The transfer between
both modes is the change of the reference voltage, so the trans-
fer between the output voltage controller and the grid current
controller does not exist in the proposed method. Four different
mode combinations with two switches have been described [23].
Switch one is to control the mode between current control and
voltage control, and the other switch is to control two operation
modes between GC mode and SA mode. The weighted param-
eter current/voltage control scheme showed good performance.
However, practically, it requires additional tuning procedures
with respect to the power rating, transfer switch delay time,
and control loop sample time. A flexible control strategy for an
11-kW wind turbine with a back-to-back power converter capa-
ble of working in both SA and GC has been proposed [24].
In this paper, we present a control method for seamless mode
transitions considering response time of mode transitions un-
der abnormal condition such as grid faults. Based on the per-
formance analysis of conventional voltage and current control
loops, a combined control strategy, which is able to overcome
the limitation of the capability of the DG inverter control un-
der mode transition conditions for unknown power plants, is
proposed in order to protect the critical load and operate the
DG inverter without grid fault trip. The proposed control loop
consists of a current controller and a feedforward voltage con-
troller, which are to minimize the grid overvoltage (OV). The
feedforward voltage control loop is added to the d–q axis current
control loop. The proposed control strategy reduces the impact
of the renewable energy and the critical load under the grid fault
or disturbance conditions.
4781
Fig. 2. Occurrence of critical load fault during conventional mode transition
under the OV condition.
The occurrence of OV in the DG inverter will be explained
in Section II. In Section III, the DG inverter controller for
operational mode transitions will be designed in GC and SA
modes. And we will investigate the response time and stability
of the conventional voltage controllers in GC and SA modes. In
Section IV, the seamless control strategy under OV conditions
will be described. The proposed control scheme will be veri-
fied by using simulation and test results in Sections V and VI,
respectively. Section VII concludes this paper.
II. IMPACT OF THE DG INVERTER UNDER OV CONDITIONS
A. Mode Transitions of the DG Inverter Under OV Conditions
Fig. 2 shows the occurrence of critical load fault during the
conventional mode transition under the OV condition. The dis-
connection time of the DG inverter from the grid according to
voltage deviations of the IEEE 1547 Standard is 1 s in 110% OV
and 0.16 s in 120% OV [25]. Therefore, the conventional mode
transition from the GC mode to the SA mode will be acceptable
during 10–60 cycles of fundamental frequency.
However, since the critical load will encounter an OV fault
due to the grid OV within one or two cycles, the DG inverter
needs to switch the operational mode immediately to avoid the
critical load fault. The grid OV results from two conditions: one
is when the PCC switch is turned OFF, and the other is when
the grid voltage is swelling.
B. Mode Transition Under the OV Condition Due to PCC
Switch-OFF
Since the DG inverter cannot recognize whether the PCC
switch is turned OFF or ON, the PCC switch can be turned
OFF before the mode switch is turned OFF when the grid is the
fault condition. At this point, the DG inverter runs with current
control in the GC mode. Then, the output voltage condition of
the DG inverter will experience the OV depending on power
conditions of the DG inverter and the critical load. The OV of
the DG inverter will be determined by
if |So | > |SL |, Sg = 0, then vo = vLim it (1)
where vLim it is the limited output value of the current controller,
So is the apparent power of the DG inverter, SL is the apparent
power in the critical load, and Sg is the apparent power in the
grid.
The output voltage of the DG inverter will reach a limit value
given in (1). It can exceed 110% of the normal grid voltage.
4782
Fig. 3. DG inverter power flow and voltage level with respect to operational
mode under the OV condition due to PCC switch-off.
Fig. 4. DG inverter power flow and voltage level with respect to operational
modes under the OV condition due to grid voltage swell.
The DG inverter will be increased up to the saturated output
voltage until the mode transition changes the controller from
the GC mode to the SA mode. After the mode transition, the
DG inverter will regulate the output voltage to the normal level
through the voltage controller in the SA mode.
Fig. 3 shows the DG inverter power flow and voltage level
with respect to operational modes under the OV condition due
to PCC switch-off. If the DG inverter can change quickly from
the GC to the SA mode within one to two cycles, then the
occurrence of the fault trip in the critical load will be reduced
because the critical load experiences less OV.
C. Mode Transition Under the OV Condition Due to Grid
Voltage Swell
Fig. 4 shows the DG inverter power flow and voltage level
with respect to operational mode under the OV condition due to
the grid voltage swell. If the DG inverter performs a fast mode
transition within one cycle under the OV condition due to the
grid voltage swell, the DG inverter fault such as an overcurrent
may occur. Although the DG inverter wants to regulate the
output voltage in the grid voltage swell, it is difficult to control
the output voltage, due to the lack of DG inverter capacity
compared to the grid. It implies that the fast mode transition in
the grid voltage swell with one or two cycles can be failed due
to the voltage control instability.
D. Requirements for Fast Mode Transition of the DG Inverter
Under OV Conditions
Fig. 5(a) shows the voltage control in the critical load during
the fast mode transition under the OV condition due to PCC
switch-off. The response time of the voltage controller depends
on the critical load variation. We assume that the minimum and
maximum critical loads are 0.5% and 50% of the rated power
of the DG. This requires the robust voltage control with one
cycle response from 0.5% critical load to 50% critical load.
Fig. 5(b) shows the voltage and current controls during the fast
mode transition under the OV condition due to the grid voltage
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 10, OCTOBER 2013
Fig. 5. Control state during fast mode transition. (a) OV condition due to PCC
switch-off. (b) OV condition due to grid voltage swell.
swell. The fast mode transition can make unexpected control
mode under the OV condition due to the grid voltage swell. The
voltage control of the DG inverter connected to the grid is not
able to regulate the grid voltage. It will cause the voltage control
instability in the worst case. Even though the DG inverter still
operates with the voltage control loop in the GC mode within
short duration, it needs to be changed from the voltage control
to the current control under the OV condition due to the grid
voltage swell. Hence, the voltage control needs to have a fast
response time under the critical load variation to order to make
the fast mode transition under the OV condition due to PCC
switch-off. In addition, the voltage control needs to guarantee
the stable operation under the OV condition due to the grid
voltage swell. Therefore, the voltage control is necessary to be
designed with considering plants in GC and SA modes. It will
satisfy a seamless control of the DG inverter with the critical
load safety under OV conditions.
III. CONVENTIONAL DG INVERTER CONTROL WITH
OPERATIONAL MODES
A. DG Inverter Modeling in GC and SA Modes
Fig. 6 shows the circuit diagram of the DG inverter with the
grid and the critical load. It consists of the three-phase dc-ac
inverter, the LCL filter, and switches. State variables are defined
as follows: v is the inverter voltage, v c is the capacitor voltage
of the LCL filter, v o is the output voltage, v g is the grid voltage,
i is the inverter current, ic is the capacitor current of the LCL
filter, io is the output current, iR is the resistive load current,
iL is the inductive load current, and iC is the capacitive load
current.
System parameters are defined as follows: Lf is the filter
inductor, Cf is the filter capacitor, Lg is the grid inductor, Rl is
the resistor of the critical load, Ll is the inductor of the critical
load, Cl is the capacitor of the critical load, and ωres is the
resonance frequency of the LCL filter. The rated capacity of the
dc–ac inverter is 30 kVA. The rated active power of the DG is
20 kW. The rated grid voltage vgrated is 440 VAC line to line.
The dc-link voltage of dc–ac inverter is 800 V. The LCL filter’
values are Lf = Lg = 2 mH and Cf = 21 μF [28], [29].
The following expression is used to transform abc axis vari-
ables to α–β axis variable into complex space vector form:
2 
xα β = xα + jxβ = xa + xb ej 2π /3 + xc ej 4π /3 (2)
3
where xα β is the complex space vector in the α–β axis.
HWANG AND PARK: SEAMLESS CONTROL STRATEGY OF A DISTRIBUTED GENERATION INVERTER FOR THE CRITICAL LOAD SAFETY 4783

Fig. 6. Circuit diagram of the DG inverter with the grid and the critical load.

To transform α–β axis variables to d–q axis variables, com- TABLE I

PARAMETERS WITH RESPECT TO POWER CONSUMPTION OF THE CRITICAL
plex space vector form is expressed as LOAD (NOMINAL LINE-TO-LINE VOLTAGE: 440 VA C , NOMINAL
xdq = xd + jxq = xα β e−j θ (3) FREQUENCY: 60 HZ)

where xdq is the complex space vector in the d–q axis. e−j θ is a
unity state vector rotating at the grid angle,θ = ωt. The d-axis is
related to the active power. The q-axis is related to the reactive
power.
From Fig. 6, complex voltage and current vector equations in
the d–q axis are given as
didq
v dq = Lf + jωLf idq + v cdq (4)
dt
dv cdq
icdq = Cf + jωCf v cdq (5)
dt
idq = iodq + icdq (6)
diodq
+ jωLg iodq + v odq .
v cdq = Lg (7)
dt
In the GC mode, the output voltage is the same as the grid
voltage
v odq = v g dq . (8)
In the SA mode, the output voltage is determined as
diL dq
v odq = Ll + jωLl iL dq (9)
dt
v odq dv odq
iodq = + Cl + jωCl v odq + iL dq . (10)
Rl dt
From (9) and (10), the complex impedance of the critical load Fig. 7. Transfer function block diagram of the DG inverter system.
is given as
v odq
= z l (s) = zla (s) + jzlb (s) . (11)
iodq 2 1
c2 = 2ω 2 + + 2 2
Impedance transfer functions are expressed as Ll Cl Rl Cl
a3 s3 + a2 s2 + a1 s + a0 2 2 2
zla (s) = (12) c1 = ω2 + , c0 = ω 4 − ω2
s4 + c3 s3 + c2 s2 + c1 s + c0 Rl Cl Rl Ll Cl2 Ll Cl
b2 s2 + b0 ω2 1
zlb (s) = (13) + + 2 2.
s4 + c3 s3 + c2 s2 + c1 s + c0 Rl2 Cl2 Ll Cl
where
1 1 ω2 1 ω2 Based on (11)–(13), we will investigate the response time
a3 = , a2 = , a1 = + , a0 = of voltage controllers in the SA mode. Parameters with respect
Cl Rl Cl2 Cl Ll Cl2 RCl2
to the critical load type are given in Table I. Fig. 7 shows the
1 ω3 ω 2 transfer function block diagram of the DG inverter system in
b2 = , b0 = − c3 =
Cl Cl Ll C2l Rl Cl either GC or SA modes based on (4)–(11).
4784
Fig. 8. Transfer function block diagram of the DG inverter in the GC mode.
Fig. 9. Transfer function block diagram of the DG inverter in the SA mode.
B. Current Control in the GC Mode
Fig. 8 shows the transfer function block diagram of the DG
inverter in the GC mode. The active damping control to remove
a resonance of the LCL filter can be added. Therefore, voltage
and current controllers are designed based on the dynamic plant
models of the DG inverter with the active damping. The d–q axis
coupling terms in Fig. 8 can be removed through feedforward
methods. The output current feedforward method in Fig. 8 will
minimize the impact of grid impedance [30]–[32].
The LCL filter with the active damping control can be ex-
pressed as a second-order system because it is damped by an
additional compensator. In Fig. 8, the v ∗cdq − v cdq transfer func-
tion, which is an equivalent active damped LCL filter, is given
as
v cdq ωL2 C
Gf (s) = = (14)
v ∗cdq s2 + 2ζωL C s + ωL2 C
where ωL C = √ 1
, ς = 0.707.
Lf Cf
Fig. 9 shows the transfer function block diagram in the SA
mode. The decoupling impedance, zlb (s) will be ignored be-
cause it is regarded as a disturbance term. Fig. 10 shows the
current control block diagram in the GC mode. The compen-
sator is designed with a proportional-integral (PI) controller.
C. Voltage Control in the SA Mode
Fig. 11 shows the voltage control block diagram in the SA
mode. The compensator is designed with a PI controller. In
the inner loop, the current control can be added. This can be
optional. Two voltage controls with/without the current loop
will have different dynamics.
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 10, OCTOBER 2013
Fig. 10. Current control block diagram in the GC mode.
Fig. 11. Voltage control block diagram in the SA mode.
D. Performance and Limitation of Voltage Control Under OV
Conditions
There are a lot of approaches to regulate the ac output voltage
of the dc–ac inverter in ac power supplies or uninterruptible
power supplies. The ac output terminal of the dc–ac inverter can
be connected to an LCL filter to reduce switching harmonics.
It can be implemented by the current control loop, in which
the inner loop is used by the current controller for overcurrent
protection and the outer loop is used by voltage controller. How-
ever, this method has the limitation in the response time of the
voltage regulation under variable and distorting loads [26], [27].
The voltage controller has to provide the fast regulation of the
output voltage during the OV condition due to PCC switch-off.
If the OV occurs by the grid swell with one cycle, the voltage
control should be stable in the grid connection. Because the
DG inverter is able to get back to the original GC mode after
recognizing the OV condition due to the grid voltage swell.
Therefore, it is important to design the voltage control with
a fast regulation and stable operation under OV conditions. It
means that the voltage control needs to consider two operational
modes. Here, we will compare the voltage control method with
the current loop and without the current loop in terms of the
response time in the SA mode and the stability in the GC mode.
Fig. 12 shows the simplified voltage control loop with the
current loop in the SA mode. The compensator is designed with
a PI controller. In the inner loop, the current loop has variable
parameters with respect to the critical load impedance zla (s).
Thus, the dynamic response of the current loop will be different
according to zla (s). Moreover, it will affect to the dynamic
response of the voltage loop. At the light-load condition, the
HWANG AND PARK: SEAMLESS CONTROL STRATEGY OF A DISTRIBUTED GENERATION INVERTER FOR THE CRITICAL LOAD SAFETY 4785

Fig. 12. Simplified voltage control loop with the current loop in the SA mode.

Fig. 15. Step responses of the voltage control without the current loop accord-
ing to the critical load type. (a) R load. (b) RL, RC, and RLC loads.

IV. SEAMLESS CONTROL STRATEGY UNDER OV CONDITIONS

A. Proposed Controller Configuration
Fig. 16 shows the proposed overall control block diagram. It
is designed based on the synchronous reference frame, which
means the d–q axis. The active power is controlled in the
Fig. 13. Step responses of the voltage control with the current loop according
to the critical load type. (a) R load. (b) RL, RC, and RLC loads.
d-axis, and the reactive power is controlled in the q-axis. In
the GC mode, the current controller with the current reference
calculation from active and reactive power references is selected
by the operational mode decision under OV conditions. In the
SA mode, the current controller is connected by the current
reference calculation with the d–q axis voltage references. The
current reference calculation will be explained in (15)–(19). In
Fig. 14. Simplified voltage control loop without the current control loop in addition, a feedforward voltage control is added in the d–q axis
the SA mode. current control. Therefore, the proposed feedforward voltage
controller in the d–q axis can help the fast voltage regulation
in the SA mode. The angle control detects the grid angle by
current will be decreased. Then, it will be more difficult to regulating the d-axis voltage.
regulate the output voltage. This controller will be worst to
regulate the output voltage at the no-load condition. B. Current Reference Calculation
Fig. 13 shows step responses of the voltage control with the The output power of the DG inverter is given as
current loop according to the critical load type in Table I. There
3
is no instability according to the critical load type. However, the Po = (vod iod + voq ioq ) (15)
response time of the current controller varies with respect to the 2
resistive, inductive, and capacitive loads shown in Fig. 13(a) and 3
Qo = (voq iod − vod ioq ) (16)
(b). Moreover, it is difficult to achieve the voltage regulation with 2
no overshoot under very light resistive load condition because where Po is the output active power and Qo is the output reactive
the inner current loop response is sensitive to the resistive load. power.
Finally, the overshoot will make additional OV. On the other From (15) and (16), the d–q axis current references with
hand, if this controller is operated under the OV condition due respect to the active and reactive power reference in the GC
to the grid voltage swell, then the DG inverter operation will be mode can be derived as
stable because the output current flow can be limited.  ∗    ∗ 
iod 2 vod −voq Po
Fig. 14 shows the simplified voltage control loop without =  2  (17)
i∗oq 3 vod + voq2 v oq v od Q ∗
o
the current loop in the SA mode. The compensator is designed
with a PI controller. Fig. 15 shows step responses of the voltage where Po∗ is the output active power reference, Q∗o is the out-
control with the current loop according to the critical load type. put reactive power reference, i∗od is the d-axis output current
Responses of the voltage control without the current loop are reference, and i∗oq is the q-axis output current reference.
fixed with respect to the resistive load shown in Fig. 15(a). The d–q axis current references with respect to the d–q axis
It implies that the voltage control without the current loop is output voltage references in the SA mode can be derived as
robust under the resistive load variation. However, it can be  ∗    ∗ 
iod 2 Po −Qo vod
unstable under either inductive or capacitive loads because there =  2  (18)
i∗oq 3 vod + voq 2 Qo P o
∗
voq
is instability shown in Fig. 15(b). Moreover, in the OV due to
∗ ∗
the grid voltage swell, it will make uncontrolled current flow, where vod is the d-axis output voltage reference and voq is the
because there is no inner current loop and current limiter. q-axis output voltage reference.
4786 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 10, OCTOBER 2013

Fig. 16. Proposed overall control block diagram.

Fig. 17. State flow diagram for operational mode decision under OV condi-
tions. (a) OV condition due to PCC switch-off. (b) OV condition due to grid
voltage swell. Fig. 18. Block diagram of the simplified current control with the feed forward
voltage loop in the SA mode.

Fig. 17 shows a state flow diagram for the operational mode

In the current loop, the current reference is calculated as
decision under OV conditions. In the GC mode, the mode switch
is turned ON. The current reference with respect to the active and i∗odq = zla
−1
(s) v ∗odq . (19)
reactive powers is determined by (17). Under the OV condition,
From Fig. 18, the output current can be expressed as
the mode switch continues to be turned ON. The current refer-
ence with respect to the active and reactive powers is changed 2
Lg ωcc
to (18). After the current reference change, the DG inverter ex- iodq = i∗
2 odq
s2
Lg + (2ςcc Lg ωcc + zla (s)) s + Lg ωcc
periences two different behaviors in the current loop during OV s ωv c
condition. +
Lg s2 + (2ςcc Lg ωcc + zla (s)) s + ωcc
2 L
g s
If the error value of the current control is decreased to near  
zero, the OV is occurred by PCC switch-off because the cur- × v ∗odq − v odq . (20)
rent reference in (18) is matched to the critical load. The DG
inverter will change from the OV condition to the SA mode in From (20), the closed transfer function in the d–q axis can be
Fig. 17(a). In the SA mode, the mode switch is turned OFF. The derived as
current reference with respect to the active and reactive powers v odq iodq ωv c s + ωcc 2
+ Lz lgaω(s)
is determined by (18). If the error value of the current control is = ∗ =   vc
(21)
v ∗odq iodq 2 z l a (s)
s + 2ζω c + L g s + ωc2 + z l a (s)
increased, the output of the current control will reach the limited Lg ωv c
value. In this case, the OV is caused by the grid voltage swell. 2
The DG inverter needs to return from the OV condition to the where kp cc = 2ςcc ωcc Lg , kicc = ωcc Lg , kpvc = ωv c /ωL C , kivc =
GC mode in Fig. 17(b). ωv c , ωcc is the bandwidth of the current controller, and ωv c is
the bandwidth of the voltage controller.
Fig. 19 shows step responses of the current control with the
C. Current Control With the Feedforward Voltage Loop
feedforward voltage loop according to the critical load type in
Fig. 18 shows the block diagram of the current control with Table I. Response times of the output voltage are almost constant
the feedforward voltage loop in the SA mode. regardless of the critical load type. It means that the proposed
HWANG AND PARK: SEAMLESS CONTROL STRATEGY OF A DISTRIBUTED GENERATION INVERTER FOR THE CRITICAL LOAD SAFETY 4787

Fig. 19. Step responses of the current control with the feedforward voltage
loop according to the critical load type. (a) R load. (b) RL, RC, and RLC loads.

Fig. 21. Simulation results of the conventional mode transition from the cur-
rent control to the voltage control without the current loop. (a) 0.1 kW resistive
load, (b) 0.1 kW resistive and 10 kVar inductive loads, (c) 0.1 kW resistive and
10 kVar capacitive loads, and (d) grid voltage swell.

Fig. 20. Simulation results of the conventional mode transition from the cur-
rent control to the voltage control with the current loop. PCC switch-off period
with (a) 0.1 kW resistive load, (b) 0.1 kW resistive and 10 kVar inductive loads,
(c) 0.1 kW resistive and 10 kVar capacitive loads, and (d) grid voltage swell.

control scheme is very robust under resistive, inductive, and

capacitive load variations. Especially, the proposed control is
able to regulate under the no-load condition because of the
feedforward voltage loop.
If this control loop is activated in the GC mode, the d–q
axis current loop is operated with the grid impedance. There
is no affect on the q-axis current loop because the q-axis feed
forward voltage control is regulated with a zero reference. On
the other hand, the d-axis feed forward voltage loop’s output is
fixed by the voltage limiter when the d-axis voltage reference
is not equal to the d-axis grid voltage. Therefore, this control
scheme is stable in the GC mode.
V. SIMULATION RESULTS
Fig. 20 shows simulation results of the conventional mode
transition from the current control to the voltage control with
the current loop. The output voltage is not regulated within one
cycle under the OV condition due to switch-off in conditions in
Fig. 20(a)–(c). Fig. 20(d) shows a stable DG inverter operation
with limited current flow.
Fig. 22. Simulation results of the proposed mode transition in the current
control with the feedforward voltage loop. (a) 0.1 kW resistive load, (b) 0.1
kW resistive and 10 kVar inductive loads, (d) 0.1 kW resistive and −10 kVar
capacitive loads, and (f) grid voltage swells.
Fig. 21 shows simulation results of the conventional mode
transition from the current control to the voltage control without
the current loop. The response time of the voltage regulation is
within one cycle under the OV condition due to switch-off period
with 0.1-kW resistive load shown in Fig. 21(a) and (b). However,
it is unstable during the mode transition with 0.1 kW resistive,
and 10 kVar capacitive loads in Fig. 21(c). Moreover, the mode
transition during grid voltage swells in Fig. 21(d) shows unstable
operation of the DG inverter in terms of uncontrolled current
flow.
4788 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 10, OCTOBER 2013

Fig. 23. Experimental setup of the hardware in the loop using a RTDS.

Fig. 22 shows simulation results of the proposed mode transi-

tion using the current control with the feedforward voltage loop.
The response time of the voltage regulation is within one cycle
under the OV condition due to switch-off period with all critical
loads in Fig. 22(a)–(c). Moreover, the mode transition during
grid voltage swells in Fig. 22(d) shows the stable DG inverter
operation with limited current flow.

VI. EXPERIMENTAL RESULTS

A. Hardware-in-the-Loop Test Using a Real-Time Digital
Simulator (RTDS)
Fig. 23 shows the experimental setup of the hardware in the
loop using the RTDS. Power switch circuits including the DG in-
verter, LCL filter, step-down transformer connecting to the main
power grid, and other power components are built in RSCAD.
They are emulated in the RTDS. The proposed seamless con-
trol algorithm was implemented by the digital controller using
the DSP control board (TMS320F28335). All pulsewidth mod-
ulated signals are transmitted to the RTDS through a GTDI card.
All voltage and current signals are measured from the output of
the GTAO card.

B. Experimental Waveforms
Fig. 24(a) shows experimental waveforms of the conventional
mode transition from the GC mode to the SA mode during PCC
switch-off period without the critical load. The control mode
of the DG inverter is changed from the current control to the
voltage control. The output voltage is regulated within 10 ms.
Fig. 24(b) shows experimental waveform of the conventional
control mode transition during grid voltage swells. The control
mode is changed from the current control in the GC mode to
the voltage control in the GC mode. After the control mode
is changed, the current of the DG inverter is increased, be-
cause the conventional voltage control has no inner current loop
with the current limiter. The conventional voltage control has
a fast regulation capability under the OV condition due to the
PCC switch-off with the no-load condition. However, if the grid
voltage swells, then the operation of the DG inverter becomes
unstable with the reverse power flow.
Fig. 25(a) shows experimental waveforms of the proposed
mode transition from the GC mode to the SA mode during
Fig. 24. Experimental waveforms of the conventional mode transition using
the voltage control without the current loop: (a) from the GC mode to the SA
mode during PCC switch-off period without critical load, and (b) from the GC
mode to the SA mode during grid voltage swells.
PCC switch-off period without the critical load. The control
mode of the DG inverter is used by the current control with
the feedforward voltage loop by changing the current reference
in the GC mode. The output voltage is regulated within 5 ms.
Fig. 25(b) shows experimental waveforms of the proposed mode
transition during grid voltage swell. The control mode of the
DG inverter is used by the current control with the feedforward
voltage loop by changing the current reference in the GC mode.
After the operational mode is changed, the limited d–q axis
HWANG AND PARK: SEAMLESS CONTROL STRATEGY OF A DISTRIBUTED GENERATION INVERTER FOR THE CRITICAL LOAD SAFETY
Fig. 25. Experimental waveforms of the proposed mode transition using the
current control with the feedforward voltage loop: (a) from the GC mode to the
SA mode during PCC switch-off period without critical load, and (b) from the
GC mode to the SA mode during grid voltage swells.
current will be generated. When the grid voltage swells, the DG
inverter will be operated by the current control with feedforward
voltage loop using the current reference in the SA mode. After
checking the grid fault, the DG inverter will change the current
reference in the SA mode to the current reference in the GC
mode without the instability shown in Fig. 25(b).
Fig. 26 shows experimental waveforms of the proposed mode
transition using the current control with the feedforward volt-
age loop. Fig. 26(a) shows the mode transition from 20 kW
active power generation in the GC mode to 0.1 kW resistive
and 10 kVar inductive loads in the SA mode. The recovery time
is within 10 ms. Fig. 26(b) shows the mode transition from
20 kW active power generation in the GC mode to 5 kW re-
sistive, 5 kVar inductive, 10 kVar capacitive loads in the SA
mode. The recovery time is within 10 ms, which means that the
proposed control method can transfer from the GC mode to the
SA mode seamlessly. Therefore, the fast mode transition can be
performed in order to minimize the OV occurrence during the
PCC switch-off period. This stable operation of the DG inverter
can be considered as a ride-through capability under the OV
condition due to the grid voltage swell.
4789
Fig. 26. Experimental waveforms of the proposed mode transition using the
current control with the feedforward voltage loop: (a) from 20 kW power gen-
eration in the GC mode to 0.1 kW resistive and 10 kVar inductive loads in the
SA mode, (b) from 20 kW power generation in the GC mode to 5 kW resistive,
5 kVar inductive, 10 kVar capacitive loads in the SA mode.
VII. CONCLUSION
This paper presented a control strategy considering the re-
sponse time of regulating output voltage in SA operational mode
transitions and the stable operation of the voltage control loop
in the grid connection mode under grid voltage abnormal con-
ditions. Conventional current control and voltage control loops
have limited response for unknown power plant conditions dur-
ing the mode transition. The proposed control loop consists of
a current controller and a feedforward voltage controller, which
are to minimize the grid OV. The feedforward voltage control
loop was added to the d–q axis current control loop. It responded
within one or two cycles to protect the critical load in the SA
operation during grid fault conditions, and regulated the output
voltage to maintain the critical load in the SA operational mode
during grid fault conditions.
The proposed control strategy reduces the impact of the re-
newable energy and the critical load under the grid fault or
disturbance conditions. In addition, the smooth operation of
the DG inverter will also enhance the stability and reliability
of the utility grid. RTDS-based hardware-in-the-loop tests and
4790
simulation results showed that the DG inverter can achieve a
seamless mode change under grid OV conditions.
REFERENCES
[1] G. O. Young, “Synthetic structure of industrial plastics (Book style with
annual energy outlook 2011),” With Projections to 2035, DOE/EIA-
0383(2011), Apr. 2011.
[2] O. C. Montero-Hernandez and P. N. Enjeti, “Ride-through for critical
loads. Exploring a low-cost approach to maintaining continuous connec-
tions between buildings and/or industrial systems,” IEEE Ind. Appl. Mag.,
vol. 8, no. 6, pp. 45–53, Nov./Dec. 2002.
[3] M. V. Aware and D. Sutanto, “SMES for protection of distributed critical
loads,” IEEE Trans. Power Del., vol. 19, no. 3, pp. 1267–1275, Jul. 2004.
[4] S. W. Mohod and M. V. Aware, “Micro wind power generator with battery
energy storage for critical load,” IEEE Syst. J., vol. 6, no. 1, pp. 118–125,
Mar. 2012.
[5] J. Kwon, S. Yoon, H. Kim, and S. Choi, “Fault ride through control
with voltage compensation capability for utility interactive inverter with
critical load,” in Proc. IEEE 8th Int. Conf. Power Electron. ECCE Asia,
2011, pp. 3041–3047.
[6] T.-S. Hwang, K.-S. Kim, and B.-K. Kwon, “Control strategy of 600 kW
E-BOP for molten carbonate fuel cell generation system,” in Proc. Int.
Conf. Electr. Mach. Syst., 2008, pp. 2366–2371.
[7] K.-S. Kim, C.-J. Lim, T.-S. Hwang, B.-K. Kwon, J.-S. Kim, C.-H. Choi,
and S.-G. Choi, “A 600KW electrical balance of plant with LCL filter
for molten carbonate fuel cell,” in Proc. 13th Eur. Conf. Power Electron.
Appl., 2009, pp. 1–7.
[8] T.-S. Hwang, H.-S. Heo, and K.-S. Kim, “High performance synchroniza-
tion algorithm for 600 kW EBOP of fuel cell generation system,” in Proc.
IEEE Int. Conf. Ind. Technol., 2010, pp. 1195–1200.
[9] E. C. Aeloiza, P. N. Enjeti, L. A. Moran, O. C. Montero-Hernandez, and
S. Kim, “Analysis and design of a new voltage sag compensator for critical
loads in electrical power distribution systems,” IEEE Trans. Ind. Electron.,
vol. 39, no. 4, pp. 1143–1150, Jul./Aug. 2003.
[10] I. J. Balaguer, Q. Lei, S. Yang, U. Supatti, and F. Z. Peng, “Control for
grid-connected and intentional islanding operations of distributed power
generation,” IEEE Trans. Ind. Electron., vol. 58, no. 1, pp. 147–157, Jan.
2011.
[11] B. Mirafzal, M. Saghaleini, and A. Kaviani, “An SVPWM-based switch-
ing pattern for stand-alone and grid- connected three-phase single-stage
boost inverters,” IEEE Trans. Power Electron., vol. 26, no. 4, pp. 1102–
1111, Apr. 2011.
[12] J. Kwon, S. Yoon, and S. Choi, “Indirect current control for seamless
transfer of three-phase utility interactive inverters,” IEEE Trans. Power
Electron, vol. 27, no. 2, pp. 773–781, Feb. 2012.
[13] Z. Yao, Z. Wang, L. Xiao, and Y. Yan, “A novel control strategy for grid-
interactive inverter in grid-connected and stand-alone modes,” in Proc.
21st Annu. IEEE Appl. Power Electron. Conf. Expo., 2006, pp. 19–23.
[14] Z. Yang, H. Liao, C. Wu, and H. Xu, “Analysis and selection of switch
for double modes inverter in micro-grid system,” in Proc. Int. Conf. Elect.
Mach. Syst., 2009, pp. 1778–1781.
[15] S. Huang, L. Kong, and H. Xu, “Control algorithm research on seamless
transfer for distributed resource with a LCL filter,” in Proc. 3rd Int. Conf.
Electric Utility Deregulation Restruct. Power Technol., 2008, pp. 2810–
2814.
[16] M. N. Arafat, S. Palle, Y. Sozer, and I. Husain, “Transition control strategy
between standalone and grid-connected operations of voltage-source in-
verters,” IEEE Trans. Ind. Appl., vol. 48, no. 5, pp. 1516–1525, Sep./Oct.
2012.
[17] G. Shen, D. Xu, and D. Xi, “Novel seamless transfer strategies for fuel
cell inverters from grid-tied mode to off-grid mode,” in Proc. 20th Annu.
IEEE Appl. Power Electron. Conf. Expo., 2005, vol. 1, pp. 109–113.
[18] Z. Liu and J. Liu, “Seamless transfer strategy with outer current loop
for three phase inverter in distributed generation,” in Proc. IEEE Energy
Convers. Congr. Expo., 2010, pp. 3556–3560.
[19] R. Tirumala, N. Mohan, and C. Henze, “Seamless transfer of grid-
connected PWM inverters between utility-interactive and stand-alone
modes,” in Proc. 17th Annu. IEEE Appl. Power Electron. Conf. Expo.,
2002, pp. 1081–1086.
[20] J. Kwon, S. Yoon, and S. Choi, “Indirect current control for seamless
transfer of three-phase utility interactive inverters,” IEEE Trans. Power
Electron., vol. 27, no. 2, pp. 773–781, Feb. 2012.
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 10, OCTOBER 2013
[21] H. Kim, T. Yu, and S. Choi, “Indirect current control algorithm for utility
interactive inverters in distributed generation systems,” IEEE Trans. Power
Electron., vol. 23, no. 3, pp. 1342–1347, May 2008.
[22] Z. Yao, L. Xiao, and Y. Yan, “Seamless transfer of single-phase grid-
interactive inverters between grid-connected and stand-alone modes,”
IEEE Trans. Power Electron., vol. 250, no. 6, pp. 1597–1603, Jun. 2010.
[23] X. Wang, C. Zhang, X. Li, and Z. Guo, “Weighted control research on
seamless transfer for dual-mode three phase inverter in micro-grid,” in
Proc. Int. Conf. Elect. Mach. Syst., Aug. 20–23, 2011, pp. 1–5.
[24] R. Teodorescu and F. Blaabjerg, “Flexible control of small wind turbines
with grid failure detection operating in stand-alone and grid-connected
mode,” IEEE Trans. Power Electron., vol. 19, no. 5, pp. 1323–1332, Sep.
2004.
[25] IEEE Standard for Interconnecting Distributed Resources With Electric
Power Systems, IEEE Standard 1547, 2003.
[26] Wasynczuk, S. D. Sudhoff, T. D. Tran, D. H. Clayton, and H. J. Hegner,
“A voltage control strategy for current-regulated PWM inverters,” IEEE
Trans. Power Electron., vol. 11, no. 1, pp. 7–15, Jan. 1996.
[27] P. Mattavelli, “An improved deadbeat control for UPS using disturbance
observers,” IEEE Trans. Ind. Electron., vol. 52, no. 1, pp. 206–212, Feb.
2005.
[28] A. A. Rockhill, M. Liserre, R. Teodorescu, and P. Rodriguez, “Grid-filter
design for a multimegawatt medium-voltage voltage-source inverter,”
IEEE Trans. Ind. Electron., vol. 58, no. 4, pp. 1205–1217, Apr. 2011.
[29] K. Jalili and S. Bernet, “Design of filters of active-front-end two-level
voltage-source converters,” IEEE Trans. Ind. Electron., vol. 56, no. 5,
pp. 1674–1689, May 2009.
[30] J. Dannehl, F. W. Fuchs, S. Hansen, and P. B. Thøgersen, “Investigation of
active damping approaches for PI-Based current control of grid-connected
pulse width modulation converters with LCL filters,” IEEE Trans. Ind.
Appl., vol. 46, no. 4, pp. 1509–1517, Jul./Aug. 2010.
[31] X. Wang, X. Ruan, S. Liu, and C. K. Tse, “Full feedforward of grid voltage
for grid-connected inverter with LCL filter to suppress current distortion
due to grid voltage harmonics,” IEEE Trans. Power Electron., vol. 25,
no. 12, pp. 3119–3127, Dec. 2010.
[32] K.-S. Kim, B.-K. Kwon, and C.-H. Choi, “A novel control algorithm of
a three-phase PWM inverter with output LC filter,” in Proc. Eur. Conf.
Power Electron. Appl., 2007, pp. 1–6.
Tai-Sik Hwang (S’11) received the B.S. and M.S.
degrees in electrical engineering from Yeungnam
University, Kyungbuk, Korea, in 2004 and 2006, re-
spectively. He is currently working toward the Ph.D.
degree at the University of Connecticut, Storrs.
In 2006, he was with Robostar Company Ltd.,
where he was involved in research on the servo mo-
tor drives for a six-axis industrial robot. In 2007, he
became an Associate Researcher at POSCON R&D
Center, Seoul, Korea, where he was involved in re-
search on the power conditioning system for molten
carbonate fuel cell generation system for two years. In 2010, he was a Research
Assistant at the Center for Clean Energy Engineering, University of Connecti-
cut. His research interests include grid-connected inverters, dc–dc converters,
electrical motor drives, battery systems, and microgrids.
Sung-Yeul Park (S’04–M’09) received the M.S. and
Ph.D. degrees in electrical and computer engineer-
ing from the Virginia Polytechnic Institute and State
University (Virginia Tech), Blacksburg, in 2004 and
2009, respectively.
From 2002 to 2004, he was a Graduate Research
Assistant at the Center for Rapid Transit Systems,
Virginia Tech, where from 2004 to 2009, he was a
Graduate Research Assistant at the Future Energy
Electronics Center. In 2009, he joined as an Assis-
tant Professor in the Department of Electrical and
Computer Engineering, University of Connecticut, Storrs, where he became an
associate member of the Center for Clean Energy Engineering. His research
interests include energy-efficient conversion, distributed generation integration,
smart building, and microgrid applications.
Dr. Park received several international paper awards including a Third Paper
Award at the IEEE Industry Applications Society in 2004, a Best Paper Award at
the Industrial Power Converter Committee 2007, an Outstanding Writing Award
at the International Future Energy Challenge in 2007, and a Torgersen Research
Excellence Award at the College of School, Virginia Tech, in 2009.