Electr Eng (2014) 96:309–321

DOI 10.1007/s00202-014-0297-y

ORIGINAL PAPER

Direct Lyapunov control (DLC) technique for distributed

generation (DG) technology
Majid Mehrasa · M. Ebrahim Adabi · Edris Pouresmaeil ·
Jafar Adabi · Bo Nørregaard Jørgensen

Received: 5 July 2013 / Accepted: 2 May 2014 / Published online: 30 May 2014
© Springer-Verlag Berlin Heidelberg 2014

Abstract In this paper, a control model based on direct of loads, under continuous injection of maximum available
Lyapunov control theory is proposed for integration of dis- active power from the DG source to the power grid.
tributed generation (DG) sources into the power grid. As a
first step, the proposed model will be elaborated in steady Keywords Distributed generation (DG) · Direct Lyapunov
state, and then proper switching state functions will be method (DLM) · Total harmonic distortion (THD) ·
defined for control of interfacing system between the DG Voltage source converter (VSC)
sources and power grid. By setting appropriate compensa-
tion current references in the control loop of the proposed
model, the active, reactive, and harmonic current components 1 Introduction
of loads will be compensated with a fast dynamic response,
thereby achieving sinusoidal grid currents in phase with load The term distributed generation (DG) refers to any electric
voltages, while required power from grid-connected load is power generation technology that is on-site or close to the
more than the maximum injected power from the DG sources load center and is integrated to the utility power grid. DG
to the grid. Using simulation, the effectiveness of the pro- sources based on renewable energy are seen as a reliable
posed control scheme is demonstrated under steady-state alternative to the traditional energy sources based on fos-
and dynamic operating conditions. The demonstration shows sil fuel [1]. The electricity business restructuring and neces-
that the proposed control model aims to: (1) uphold a unity sity of producing more electrical energy, combined with the
value for the power factor of the grid by injection of reactive environmental regulations due to greenhouse gas emission,
power; and (2) reducing the harmonic current distortion of and the recent improvement in small-scale sources of power
the grid current by injection of harmonic current components generation are the main factors driving the energy sector
into a new era of power generation, where large portions
M. Mehrasa of increases in electrical energy demand will be met through
Young Researchers and Elite Club, Sari Branch, Islamic Azad widespread installation of DG sources [2]. Rural and house-
University, Sari, Iran hold consumers are the cases which are mostly concerned
M. Ebrahim Adabi
with the improvement of DG technology because of the over-
Faculty of Electrical Engineering, Center of Excellence for Power whelming investments required to connect to a distant trans-
System Automation and Operation, Iran University of Science and mission line. For these consumers, use of DG technology
Technology, Tehran, Iran especially based on renewable energy sources (RES) is more
E. Pouresmaeil (B)· B. N. Jørgensen
economical and efficient than the central power station sys-
Faculty of Engineering, Centre for Smart Energy Solutions, tem plus associated transmission and distribution line expan-
University of Southern Denmark (SDU), Odense, Denmark sion [3]. On the other hand, because of high reliability and
e-mail: [email protected] low cost effectiveness of DG technology, many industrial
J. Adabi
companies, and commercial consumers may decide to install
Faculty of Electrical and Computer Engineering, Babol Noshirvani DG system as a match with the electric main source for their
University of Technology, Babol, Iran electrical consumptions [4]. Moreover, a DG system could

123
310 Electr Eng (2014) 96:309–321

appear as an independent power system that meets both the tem is investigated asymptotically by definition of positive
local loads and main grid requirements, such as injection of certain scalar function based on d–q components of injected
active power into the power grid, compensation of higher current from DG. Then, switching state functions of inter-
harmonic components and reactive power of grid-connected faced VSC are achieved in dynamic condition by consider-
loads, power factor correction of main grid, backup genera- ing disturbances caused by parameters uncertainty and load
tion during overload condition, compensation of power qual- changes. The effectiveness of the proposed control technique
ity events during disturbances, peak shaving, and voltage reli- is validated by obtaining sinusoidal voltage at the point of
ability enhancement in a right way that is not possible with common coupling (PCC), lower THD for the grid current,
centralized generation [5,6]. But, increasing number of DG and active and reactive power control, in both the dynamic
units in the power grid requires new techniques for operation and steady-state operating conditions.
and management, in order to maintain or even to improve
the power supply reliability and quality in future. Conse-
quently, design of a control technique, which considers dif- 2 The proposed DG model analysis
ferent aspects of the power grid, becomes of high interest in
DG technology. 2.1 Grid-connected DG model
Several control techniques have been proposed for integra-
tion of DG units into the power grid [7–11]. In [12] impacts Figure 1 shows the schematic diagram of the proposed DG
of DG on voltage control, grid protection, and fault levels are model. Conventional signs of the variables are illustrated
investigated. A PLL-less control technique for integration of in this figure, where RDG and L DG represent the equiv-
DG units to the grid is presented in [13]. Due to sensitiv- alent resistance and inductance of DG, Rg and L g rep-
ity of PLL to noises and distortion, its elimination can bring resent the grid resistance and inductance up to the PCC;
benefits for robust control against distortions in DG applica- VPCCk (k = a, b, c) is the supply voltage components at the
tions. A delay compensation method is presented in [14] by PCC; Vgk is the grid voltage components; Vdc is the dc-link
employing feedback from the summation of current errors. voltage; and i gk , ilk and i DGk are grid, load, and DG cur-
The main weakness of this method is its large overshoot and rent components, respectively. In addition, the DG source
long transient period during integration of DG into the grid. and additional components are represented as a dc current
An overview of different control techniques for DG sources source, which is connected to the dc side of the interfaced
in the power grid is presented in [15]. This overview discusses converter.
different hardware structures, control strategies for the grid- By connection of nonlinear loads to the grid, nonlinear
side converter, and control strategies under fault conditions. currents are drawn from the grid and the grid will be pol-
In [16], different implementation techniques for DG system luted by the harmonic current components. In addition, the
are presented, and their major characteristics are pointed out. power factor of the grid will be reduced by connection of
A discussion about different controllers in DG and their abil- reactive loads to the grid. Although, the main objective of
ity to compensate low-order harmonic components presented DG technology is to provide active power for loads; nonethe-
in the main grid is presented in [17]. In [18], an overview of less, by means of a control technique for power electronic
grid synchronization strategies, their influences and roles in devices, reactive power and harmonic current components
the control of DG system on normal and faulty grid con- can be compensated and the DG model can hereby be used
ditions were discussed. A DLM control technique has been as power-factor-correction and active-power-filter devices.
proposed for active and reactive power control and stabil-
ity of control system in a single-phase grid-connected PV 2.2 Dynamic model analysis of the proposed DG model
inverter in [19]. In [20,21], a control technique has been sug-
gested to decrease harmonics of grid currents and control To draw an appropriate plan to control the integration of DG
active and reactive power of the grid through a three-phase sources to the power grid, a dynamic analytical model of
grid connected PV inverter for micro-grid applications. In the proposed DG model should be developed. According to
[21], spatial repetitive controller (SRC) is used to improve Fig. 1, the Kirchoff’s laws for voltages and currents applied
current controller implementation for estimation of periodic at PCC leads to,
disturbance of the system.
This paper presents a direct Lyapunov method (DLM) to di DGk
L DG + RDG i DGk + VPCCk + Vkn = 0 Vkn
control a grid-connected DG model. Dynamic and steady- dt
state model of the proposed DG model will be analyzed in = Vko − Vno k = a, b, c
⎛ ⎞ (1)
detail. Switching functions of the interfaced converter will
dVdc 1 
b.c 
b.c
be established by investigation of the proposed DG model in Cdc =− ⎝ Sk − 1 S j ⎠ i k + Idc ,
steady-state operating conditions. Then, stability of the sys- dt 2 3
k=a j=a

123
Electr Eng (2014) 96:309–321 311

Three-Phase Grid
Battery
n Three-Phase Nonlinear Load1

Wind Pg + jQg ig a ig b ig c
+ ila1
PV
Three-Phase Voltage Source Inverter Vga Vgb −Vg ilb1
c ilc1 R
Rg L1

ic S Sb Sc sw Lg
I dc Vdc + a
RDG LDG
2C dc
2 − a
O iDG b
Vdc + a iDG c
2C b
iDG ila 2
2 − dc c Point of Common
Coupling ilb 2
Sa Sb Sc ilc 2 R
L2

PDG + jQDG
Pl + jQl Three-Phase Nonlinear Load2

Fig. 1 Functional diagram of the proposed model

where the switching function of Sk can be defined as: By substituting (5) in (4), (6) can be expressed as:
 di DGa
1, if Tk is on and Tk is off L DG + RDG i DGa + Vdc u eqa + VPCCa = 0
Sk = (2) dt
0, if Tk is off and Tk is on
di DGb
L DG + RDG i DGb + Vdc u eqb + VPCCb = 0
According to Fig. 1, equations between switching func- dt (6)
di DGc
tions of VSC and phase differences between the point n and L DG + RDG i DGc + Vdc u eqc + VPCCc = 0
phases a, b, and c can be defined as: dt
dVdc
⎡ ⎤ ⎡ ⎤⎡ ⎤ Cdc − u eqa i DGa − u eqb i DGb − u eqc i DGc − i dc = 0
Vcn 2 −1 −1 Sa dt
⎣ Vbn ⎦ = 1 × Vdc ⎣ −1 2 −1 ⎦ ⎣ Sb ⎦ (3) Equation (6) verifies dynamic equations of the proposed
3 2
Vcn −1 −1 2 Sc DG model in abc frame that completely describes the behav-
ior of the proposed DG model.
By substituting (3) in (1), (4) can be expressed as:
⎛ ⎞ 2.3 Steady-state analysis of the proposed DG model
1
b,c
di DGk 1
L DG + RDG i DGk − ⎝ Sk − S j ⎠ Vdc In order to facilitate the performance of the proposed con-
dt 2 3
j=a trol strategy, the model can be transformed to the synchro-
+VPCCk = 0 k = a, b, c (4) nous orthogonal frame, which rotating at the grid angular
⎛ ⎞
frequency, ω, where all AC variables become DC, then con-
1 ⎝ 1 ⎠
b,c b,c
dVdc
Cdc =− Sk − S j i k + i dc trolling and filtering of the proposed model will be achieved
dt 2 3 easier. By applying the park transformation matrix to (6),
k=a j=a
dynamic equations can be expressed in d-q frame as:
In which Sk represents switching function of VSC in phase
di DGd
k. To simplify the dynamic equation of DG model, equivalent L DG + RDG i DGd −ωL DG i DGq +Vdc u eqd
dt
switching functions u eqk are defined as,
+VPCCd = 0
⎧ di DGq
⎪
⎪ u eq = − 2Sa −S6 b −Sc L DG + RDG i DGq +ωL DG i DGd +Vdc u eqq (7)
⎨ a dt
u eqb = − 2Sb −S6 a −Sc (5) +VPCCq = 0
⎪
⎪
⎩ dVdc
u eqc = − 2Sc −S6a −Sb Cdc − u eqd i DGd − u eqq i DGq − i dc = 0
dt
As can be seen from (5), equivalent switching functions The proposed model in Fig. 1 must be controlled in order
depend on all three switching functions of interfaced VSC. to generate the required active and reactive power for non-

123
312 Electr Eng (2014) 96:309–321

I refq (A)
linear loads at different frequencies during the steady-state I DGq max
condition. DG should provide the maximum possible share Consumption area of
Load Current
of load active power and all the reactive power at fundamen- ilqmax
tal frequency. Therefore, voltages of PCC in steady states can
be expressed as:

VPCCd = VPCCm , VPCCq = 0 (8)

−V pcc
2

C m
,0
Assuming that the desired value for the DG current to 2 RDG ildmax I refd (A)
I DGd min
provide active current at steady state is equal to Iref d ; then, *
V I V 2
pccm
I DGds max
+
dc dc
in a rotating reference synchronous frame, where the q- RDG 4R 2
DG

component of the load current is perpendicular to the volt-

age’s d-component, the value of the current’s q-component
to provide the required reactive power of the load, will be
equal to Iref q ,
I DGq min
i DGd = Iref d (9)
i DGq = Iref q (10) Fig. 2 i ref q versus i ref d in proposed DG model

Equations (8) to (10) are considered as stability criteria.

In steady state, Vdc is equal to its desired value Vr . If d and q In stable condition ilq = i DGq , therefore,
components of switching functions are assumed as u eqds and 

u eqqs in steady state, and substituting (8) to (10) in (7) and  Vr Idc 2
VPCC VPCCm
Vdc → Vr , (11) can be expressed as: 0 ≤ i DGd ≤  + 2
m
− il2q − (17)
RDG 4RDG 2RDG
RDG Iref d − ωL DG Iref q + Vr u eqds + VPCCm = 0 Equation (17) shows the criteria of i DGd for a stable operation
RDG Ir e fq + ωL DG Ir e fd + Vr u eqqs = 0 (11) of the proposed control plan. As can be seen from Fig. 2,
injected currents from DG can follow d and q components of
u eqds Iref d + u eqqs Iref q + Idc = 0
load currents. According to Fig. 1, VoutDG = VPCCd + j VPCCq
According to (11), switching state functions of interfaced and IoutDG = i DGd + ji DGq ; therefore, SDG can be expressed
VSC can be calculated in the steady-state operating condition as:
as follows: ∗
SDG = VoutDG Iout DG
= PDG + j Q DG (18)
−VPCCm − RDG Iref d + ωL DG Iref q
u eqds = (12) By applying stability conditions to Eq. (18),
Vr
−RDG Iref q − ωL DG Iref d PDG = VPCCm Iref d (19)
u eqqs = (13)
Vr
Q DG = −VPCCm Iref q (20)
By substituting (12) and (13) in last part of (11)
Equations (19) and (20) show that active and reactive power
2
RDG Iref d
+ RDG Iref
2
q
+ VPCCm Iref d − Vr Idc = 0 (14) of DG can be controlled by Ir e fd and Ir e fq , respectively. By
substituting (19) and (20) in (15), (21) can be expressed as:
Equation (14) can be rewritten as:
 2
2  2

2 2 VPCC Vr Idc VPCC
VPCCm Vr Idc VPCC PDG + m
+ Q DG = VPCCm
2 2
+ m
Iref d + + Iref
2
q
= + 2
m
(15) 2RDG RDG 4RDG 2
2RDG RDG 4RDG
(21)
The trajectories of the DG currents in d-q representation are
depicted in Fig. 2. As shown in this figure,(15) is a circle Equation (21) can be shown by a circle with center of
  −VPCC
2
Vr Idc
2
VPCC
, 0 and radius of VPCCm +
2
−V m Vr Idc VPCC m m
. The tra-
with center of 2RPCC DG
, 0 and radius of R DG
+ 4R 2
m
, 2RDG RDG 2
4RDG
DG
which determines the maximum and minimum value of i DGd jectories of the DG active and reactive power are shown in
and i DGq in the steady-state operating condition. For effective Fig. 3. This figure is the capacity curve (CC) of the proposed
operation of the proposed DG model we shall have, DG model, which shows the capability of the proposed DG
model in generating active and reactive power. As shown
IDGq min ≤ ilq ≤ IDGq max (16) in Fig. 4, the proposed DG model can supply all the active

123
Electr Eng (2014) 96:309–321 313

and reactive power of Load 1, but by adding Load 2, total Q l2 max = Q DG2 (22)
required power from load is more than maximum power of Pl2 max = PDG2 max + Pg (23)
DG. Therefore, some part of active power in fundamental
frequency should be supplied by utility grid. where Pg is the active power that is injected from the power
Load increment in Fig. 4 is in a way that the total reactive grid to the load. If we consider the Pg in the fundamental
power and a part of active power are supplied by DG, which frequency, DG can be also used for Load 2 and the grid. In
means, order to make a proper operation the proposed DG model,
PDG should be achieved as:
  

 Vr Idc
2
VPCC 2
VPCC
QDG (kVar)

PDG = VPCCm 2 + m
− Q 2 − m
2 l2 max
RDG 4RDG 2RDG
(24)
Figure 5 indicates CC of different DG along with con-
sumption area of typical loads. According to Fig. 5, CC1
is not suitable for consumption load. Power factor, voltage
2
V pcc quality and capability of DG in generating active and reactive
C m
,0 power increase noticeably when the radius of CC increases.
2 RDG
PDG (kW) As can be seen from this figure, DG with capability curve of
CC2 and CC3 are suitable for supplying related loads. CC2
Vr I dc
2
V pcc shows after connection of second load to the grid, DG sup-
V pccm m

RDG 2
4 RDG plies the entire loads and injected power from the grid side
to the load is zero; but, CC3 shows DG can supply the entire
load and also can injects some power to the power grid.

3 DLM control technique representation

DLM method investigates stability of grid-connected DG

Fig. 3 CC of the proposed DG model model and approaches state variables to equilibrium points

Operating area of
QDG (kVar) Load2

Ql 2max

Ql1max

Region 1 Region 2
Operating area of
Load1

PDG (kW)

2 Pl1max Pl 2 max
V pcc Pg
V 2 C m
,0
Vr I dc pccm 2 RDG
V pccm 2
RDG 4R DG

Fig. 4 Consumption area for the Load 1 and Load 2 in CC, while load power is more than the maximum power of DG

123
314 Electr Eng (2014) 96:309–321

QDG (kVar)

Qlmax

R3 Operating area of
typical Load
R2
R1

PDG (kW)

C3 C2 C1 Plmax

CC1

CC 2

CC 3

Fig. 5 CC of DG with different capacities

by considering large signal disturbances in direction of these to achieve system stability,

points. The proposed DLM strategy is a nonlinear controller
without linearization of DG nonlinear equations, which leads Ḣ (x1 , x2 ) = L DG ẋ1 x1 + L DG ẋ2 x2 0 (28)
to an asymptotical stability of the proposed model.
Equation (28) clarifies that DG model state variables fol-
3.1 Dynamic state analysis of the proposed DG model
low their equilibrium points. In order to guarantee the stabil-
ity of DG and increase its robustness against undesirable dis-
In a real system, disturbances will appear due to the inher-
turbances, the switching function of the converter in dynamic
ent uncertainty of the parameter and some abrupt changes
and steady states can be defined as:
of loads. Therefore, a control technique based on DLM is
employed to achieve proper dynamic response of DG link
during these unexpected changes. The following positive cer- u eqd = u eqds + u eqd (29)
tain scalar function is determined to validate the stability of u eqq = u eqqs + u eqq (30)
the proposed DG model.
1 1 The u eqd and u eqq show the flexibility of the proposed
H (x1 , x2 ) = L DG x12 + L DG x22 (25) controller against disturbances. With respect to (28) and the
2 2
first part of (7),
In the above equation input DC voltage is considered constant
with Vdc = Vr . Two state variables x1 and x2 are defined as:
L DG ẋ1 = −RDG i DGd +ωL DG i DGq −Vr
x1 = i DGd − Ir e fd (26) ωL DG Ir e fq − RDG Iref d −VPCCm
x2 = i DGq − Ir e fq (27) ×
Vr
Equation (25) is the total energy of whole system in the con- − Vr u eqd −VPCCd + (RDG Iref d − ωL DG Iref q
trol region which should be dissipated completely. Therefore, + Vr u eqds + VPCCm ) (31)

123
Electr Eng (2014) 96:309–321 315

Similar calculations can be used for L DG ẋ2 , u eqd = γ (i DGd Vr − Iref d Vr ) (36)
u eqq = λ(i DGq Vr − Iref q Vr ) (37)
L DG ẋ2 = −RDG i DGq − ωL DG i DGd − Vr
−RDG Iref q −ωL DG Iref d where, γ and λ are constant coefficients of the switching
× −Vr u eqq − VPCCq state functions during dynamic states. The proposed Lya-
Vr
+(RDG Iref q + ωL DG Iref d + Vr u eqqs ) (32) punov model, which is a combination of (12), (13), (36) and
(37) creates an asymptotical stable region for DG operation
By multiplying (28) in (33) and (29) in (34), in which d and q components of DG current can precisely
track their reference values in both the steady and dynamic
L DG ẋ1 x1 = −RDG (i DGd − Iref d )2 operating states. Figure 6 shows the general structure of the
+ωL DG (i DGd − Iref d )(i DGq − Iref q ) proposed DLM technique.

−Vr u eqd (i DGd − Iref d ) (33)

3.2 Reference current calculation for the proposed DG
L DG ẋ2 x2 = −RDG (i DGq − Iref q ) 2
model
−ωL DG (i DGq − Iref q )(i DGd − Iref d )
−Vr u eqq (i DGq − Iref q ) (34) It is necessary to define suitable reference values in the cur-
rent control loop for effective operation of DG unit during
By substituting Eq. (33) and (34) in the Eq. (28), dynamic and steady-state operating conditions. The total har-
monic generated by nonlinear load in steady state and during
Ḣ (x1 , x2 ) = −RDG (i DGd − Iref d )2 load change should be considered in order to calculate suit-
−RDG (i DGq − Iref q )2 able values for the reference parameters. If the total harmonic
current components of nonlinear loads supplied by DG, the
−u eqd (Vr i DGd − Vr Iref d )
injected current from the power grid will be free of harmonic
−u eqq (Vr i DGq − Vr Iref q ) (35) current components. On the other hand, the maximum capac-
ity of VSC for injection of active and reactive power at the
According to (35), switching state functions of the inter-
main frequency from DG to the grid should be considered.
faced converter can be defined as follows to have positive
As can be seen from (19), active power can be controlled by
definite total energy of DG and tolerate against dynamic
d-component of DG current (i DGd ) and d-component of load
changes,
currents can be defined as:
∞

ild = ildh1 + ildh(2k+1), (38)
DG System
k=1
Vdc
2 Grid+3phase nonlinear where ildh1 is the d-component
 of the load current at the
main frequency and ∞
O
k=1 ldh(2k+1) is the harmonic current
i
loads
Vdc
2 components of load currents in d-axis. By means of a low-
pass filter (LPF) total harmonics of ild can be expressed as:
iDGd iDGq ∞

References
iDGd iDGq ildh(2k+1) = ild (1 − LPF) (39)
Generator k=1
I refq I refd I refd I refq

Lyapunov Based Controller Therefore, DG must inject the ∞ k=1 ildh(2k+1) to the grid to
Dynamic-State Steady-State supply the harmonic current components of load currents.
Analysis of DG Analysis of DG The reference current value of DG control loop for injecting
System System
the maximum active power at fundamental frequency from
DG link to the grid can be obtained as:
Eq.(37) Eq.(36) Eq.(12) Eq.(13)
PDGmax
i DGd max = (40)
VPCCm
where, PDGmax is the maximum active power of DG at the
uedd
u ed q fundamental frequency. According to (39) and (40), the ref-
∗ ) for injecting
erence current value of DG control loop (i DG d
Fig. 6 Lyapunov-based control diagram for the proposed DG model the total harmonic current components in d-axis and a maxi-

123
316 Electr Eng (2014) 96:309–321

PDGmax

Maximum active power of DG in fundamental

frequency
V pccm
ild I refd
Low Pass PI Controller
Filter (Kpd+KId/S)

iDGd
(a)

I refq
iDGq PI Controller
(Kpq+KIq/S)

ilq
(b)
Fig. 7 Reference current components of DG control loop: a d-axis, b q-component axis

mum active power in the fundamental frequency (PDGα ) can cates the reference currents generating process in the control
be expressed as: loop of DG unit.
 ∞
∗ PDGα PDGα
i DG d
= + ildhn = +ild (1 − LPF), (41)
VPCCm VPCCm 4 Results and discussion
n=2

where
The simulation results are carried out with MATLAB/
PDGα = α PDGmax 0 ≤ α ≤ 1 (42) SIMULINK to verify the performance of proposed control
method for integration of DG sources to the power grid.
The reference current of d-axis in the proposed DG control General schematic diagram of the proposed model with con-
loop can be expressed as: trol technique, and proposed model parameters are shown
K Id in Fig. 8 and Table 1, respectively. A power converter with
∗
Ir e fd = (i DGd − i DG d
) K pd + , (43) the power rating of 17.5 kVA is utilized in this model. The
S
main objective of this section is to investigate the capabil-
where K pd and K Id are the proportional and integral gains ity, flexibility and robustness of the proposed control tech-
of the d-component of PI controllers, respectively. In an nique in controlling the interfaced converter to generate
ideal system, total reactive power of loads should be sup- required active and reactive power in both the steady-state
plied through the DG link. By assuming that reactive power and dynamic operating conditions. Harmonic current com-
is in the range of DG capacity, pensation, power factor correction and DG sources intercon-
nection with the grid by abrupt nonlinear load changes are
Q DG = Q l (44)
also examined.
In order to compensate for the reactive power of loads at
fundamental and harmonic frequencies, DG must inject ilq . 4.1 DG connection to the grid with nonlinear load
For this purpose, increment
Ql Q DG
i lq = = = i DGq (45) Before connection of DG link to the grid, a nonlinear load
−VPCCm −VPCCm
with Pl = 6.5 kW and Q l = 3.5 kVar at the fundamental
Therefore, the q-component of the reference current in frequency are supplied through the power grid. This load
q-axis can be expressed as: draws the harmonic currents from the grid, and this process
K Iq is continued until, t = 0.1 s, while DG is connected to the
Iref q = (i DGq − ilq ) K pq + (46) grid. At t = 0.2 s, another load is added to the previous
S
load and total loads increased to Pl = 15 kW and Q l =
where, K pq and K Iq are proportional and integral gains of 9 kVar. Figure 9 shows the load voltage and grid current in
the q-component of PI controllers, respectively. Figure 7 indi- phase ‘a’ before and after connection of DG to the grid, and

123
Electr Eng (2014) 96:309–321 317

n
Increasing and
decreasing RL
Three-Pahse Grid aditional Load
with Breaker
Idc ig a
ig b
VSI sw
iDGa LDG ig c ila
Vdc RDG V pcca
2 iDGb RDG LDG V pccb ilb
SPWM O
iDGc RDG LDG V pccc ilc RL
Vdc
2
Proposed Controller
odq/abc abc/odq
Lyapunov-Based Controller abc/odq abc/odq
iDGd iDGq
ueqq ueqd Eq(37)
Dynamic-State V pccd V pccq PDGmax
Analysis of DG
ilq ild
Eq(36) System

Eq(12) Maximum Active Power

Steady-State
of DG in f=50Hz
Analysis of DG
Eq(13) System V pcc PDGmax
d

I refq I refd PI Controller

RDG LDG Vdc (Kpd+KId/S)
Low Pass Filter

System Parameters PI Controller

(Kpq+KIq/S)
References Generator

Fig. 8 Overall diagram of the proposed model included by DLM

Table 1 Simulation parameters total reactive power is supplied by DG. Additional load is
Grid voltage 380 rms V disconnected from the grid at t = 0.3 s. Figure 9b shows
Input voltage 800 V DC the ability of the proposed control strategy to track the load
Main frequency 50 Hz current, when a abrupt change occurs in load currents.
Inverter resistance 0.1  Figure 10 shows the load, grid and DG currents in phase
Inverter inductance 0.45 mH ‘a’. As can be seen from this figure, DG current becomes
Coefficient of γ 0.001 equal to the load current after connection of DG to the grid.
Coefficient of λ 0.0001 After adding additional load to the grid, grid currents become
Switching frequency 10 kHz sinusoidal and after removing this additional load, the whole
PDGmax 6.5 kW
load current is supplied through the DG link.
Figures 11 and 12 illustrate the ability of control loop to
track the reference current trajectories of d- and q-axes, dur-
during load increment. As depicted in Fig. 9a, nonlinear load ing connection of DG to the grid and additional load incre-
is supplied by the utility grid before integration of DG link ment, respectively. Figures 11a and 12a show that, before
to the grid. After connection of DG to the grid at t = 0.1 s, connection of additional load to the grid, the actual d- and q-
injected current from grid reaches zero without any transient axis current components of DG’s control loop track their ref-
and the whole load current is supplied by DG. Figure 9b erence trajectory precisely. However, according to Figs. 11b
shows that the grid current increases after load increment at and 12b, after connection of additional load to the grid, the
t = 0.2 s, because the required power of the loads exceeds actual d-axis current component of DG tracks half of its ref-
the maximum injected power from DG; therefore, the rest of erence trajectory which is equal to its maximum active power
the required power is injected by the power grid. As shown (reference active power), and all the reference trajectories of
in this figure, grid current is in phase with load voltage and reactive current change.

123
318 Electr Eng (2014) 96:309–321

Fig. 9 Load voltage and grid current: a before and after connection of Fig. 11 d-component of load and DG currents: a before and after con-
DG to the grid, b before and after additional load decrement nection of DG link to the grid, b before and after additional load decre-
ment

Fig. 10 Load, DG, and grid currents of phase “a”: a before and after Fig. 12 q-component of load and DG currents: a before and after con-
connection of DG to grid, b before and after additional load decrement nection of DG to grid, b before and after additional load decrement

Figure 13 shows the contribution of DG and grid in sup- does not exist, load active power is supplied through the DG.
plying load active power before and after integration of DG In the time interval of 0.2 < t < 0.3, Pg = 6.5 Kw and
link to the grid, and before and after additional load decre- total harmonics are supplied by DG. Figure 14 shows reac-
ment. As can be seen from this figure, when additional load tive power of DG, load and grid before and after integra-

123
Electr Eng (2014) 96:309–321 319

Fundamental(50Hz)=38.5(A)
(a) 40

32
THD 16.5%

Amplitude
24

16

8
0 0 0
0
1 3 5 7 9 11 13 15 17
Harmonics Order

Fundamental(50Hz)=18.7(A)
(b) 25

20

Amplitude
THD 3.01%
15

10

5
0 0 0
Fig. 13 Active power sharing between DG, load, and grid: a before 0
and after connection of DG to grid, b before and after additional load 1 3 5 7 9 11 13 15 17
decrement Harmonics Order

Fig. 15 Current spectra: a load current spectra during connection of

additional load to the grid, b grid current spectra during connection of
DG to the grid after connection of additional load

in fundamental and harmonic frequencies are supplied by

DG.

4.2 Analysis of THD and power factor of utility grid

The spectra of load and grid currents are shown in Fig. 15

after connection of additional load to the grid, to evaluate the
capability of proposed control technique to compensate the
harmonic current components of grid currents during con-
nection of nonlinear loads to the grid. Figure 15a shows the
THD of load current in phase (a) during connection of addi-
tional load to the grid. As indicated in this figure, THD of
load current is 14.5 % and it can be considered as the THD of
grid current without DG connection. Figure 15b shows the
THD of grid current during connection of DG to the grid,
and as shown in this figure THD of grid current reduced
from 16.5 to 3.01 %, which validate the proper performance
of proposed control strategy in harmonic compensation of
Fig. 14 Reactive power sharing between DG, load, and grid: a before
and after connection of DG to grid, b before and after additional load nonlinear grid-connected loads.
decrement Table 2 shows the harmonics of grid currents in three
phases after additional load increment. As it can be seen from
tion of DG to the grid, and before and after additional load this table, THD of i ga , i gb and i gc are reduced from 16.5, 16.42
decrement. As depicted in this figure, all the reactive power and 16.45 % to 3.01, 3.03 and 3.02 % during integration of

123
320 Electr Eng (2014) 96:309–321

Table 2 Total harmonic distortion of grid currents into the power grid. High performance of the proposed con-
Grid currents THD 380 rms V trol technique in both steady-state and dynamic operating
conditions has been verified through the simulation results.
BC (%) AF with increment (%)
When using the proposed control technique, problems due
i grida 16.5 3.01 to synchronization of DG sources to the power grid will not
i gridb 16.42 3.03 exist; therefore, DG can be integrated with the power grid
i gridc 16.45 3.02 without the risk of current overshoot. Another advantage of
the proposed method is its fast dynamic response in track-
ing reactive power variations. By using the proposed control
method, DG technology is introduced as a new alternative for
distributed static compensator in utility grid. The obtained
results illustrated that under all conditions load voltage and
grid current are in phase; hence, by improvement of power
factor at PCC, DG can act as a power factor corrector device.
The results indicate that DG can provide the required har-
monic current components of nonlinear load; thus, by reduc-
ing THD of grid current, it can act as an active power filter.
The proposed control technique can be used for different
types of DG sources as power quality improvement devices
in a utility power grid.
Fig. 16 Power factor between the load and grid currents, and load
voltage

Table 3 Power factor correction

References
Power factor PF 380 rms V

BC (%) AF with increment (%) 1. Pepermans G, Driesen J, Haeseldonckx D, Belmans R, D’haeseleer

W (2005) Distributed generation: definition, benefits and issues.
PF1 87.94 100 Energy Policy 33(6):787–798
2. Rahimi F, Ipakchi A (2010) Demand response as a market resource
PF2 86 100
under the smart grid paradigm. IEEE Trans Smart Grid 1(1):82–88
PF3 87.1 100 3. Akorede MF, Hizam H, Pouresmaeil E (2010) Distributed energy
resources and benefits to the environment. Renew Sustain Energy
Rev 14(2):724–734
4. Farret FA, Simoes M (2006) Integration of alternative sources of
DG into the power grid. Figure 16 shows the power factor energy. IEEE Press, Wiley, Hoboken, New Jersey
between load and grid currents and load voltage in phase (a) 5. Naderi S, Pouresmaeil E, Gao DW (2012) The frequency-
during connection of DG to the grid and after additional load independent control method for distributed generation systems.
increment. As can be seen from this figure, the power factor Appl Energy 96:272–280
6. Pouresmaeil E, Montesinos-Miracle D, Gomis-Bellmunt O,
of the load current and load voltage is 87.94 % which can Bergas-Jan J (2010) A multi-objective control strategy for grid
be considered as the power factor between grid current and connection of DG (distributed generation) resources. Energy
load voltage without connection of DG to the grid. But, after 35(12):5022–5030
connection of DG to the grid, power factor between grid cur- 7. Karimi H, Yazdani A, Iravani R (2011) Robust control of an
autonomous four-wire electronically-coupled distributed genera-
rent and load voltage reaches 100 % during additional load tion unit. IEEE Trans Power Deliv 26(1):455–466
increment. The same scenario is done for the other phases 8. Cardenas A, Guzman C, Agbossou K (2012) Development of a
as well. Table 3 shows the power factor values between grid FPGA based real-time power analysis and control for distributed
currents and load voltages during connection of DG to the generation interface. IEEE Trans Power Syst 27(3):1343–1353
9. Pouresmaeil E, Montesinos-Miracle D, Gomis-Bellmunt O (2012)
grid and additional load increment. This performance indi- Control scheme of three-level inverter for integration of renewable
cates the capability of DLM in DG technology as a power energy resources into AC grid. IEEE Syst J 6(2):242–253
factor correction device; then, no reactive power is generated 10. Pouresmaeil E, Gomis-Bellmunt O, Montesinos-Miracle D,
by utility grid to supply the reactive power of loads. Bergas-Jané J (2011) Multilevel converters control for renewable
energy integration to the power grid. Energy 36(2):950–963
11. Mehrasa M, Adabi ME, Pouresmaeil E, Adabi J (2014) Passivity-
5 Conclusion based control technique for integration of DG resources into the
power grid. Int J Electr Power Energy Syst 58:281–290
12. Coster EJ, Myrzik JMA, Kruimer B, Kling WL (2011) Integration
A flexible control scheme based on direct Lyapunov theory issues of distributed generation in distribution grids. Proc IEEE
has been proposed in this paper for integration of DG sources 99(1):28–39

123
Electr Eng (2014) 96:309–321 321

13. Pouresmaeil E, Miguel-Espinar C, Massot-Campos M, 19. Dasgupta S, Sahoo SK, Panda SK (2010) A novel current control
Montesinos-Miracle D, Gomis-Bellmunt O (2013) A control scheme using Lyapunov function to control the active and reactive
technique for integration of DG units to the electrical networks. power flow in a single phase hybrid PV inverter system connected
IEEE Trans Ind Electron 60(7):1–13 to the grid. In: Conf. Rec. IEEE IPEC, pp 1701–1708
14. Kukrer O (1996) Discrete-time current control of voltage-fed three- 20. Dasgupta S, Mohan SN, Sahoo SK, Panda SK (2011) A Lyapunov
phase PWM inverters. IEEE Trans Power Electron 11(2):260–269 function based current controller to control active and reactive
15. Blaabjerg F, Teodorescu R, Liserre M, Timbus AV (2006) Overview power flow in a three phase grid connected PV inverter under gen-
of control and grid synchronization for distributed power genera- eralized grid voltage conditions. In: 8th international conference
tion systems. IEEE Trans Power Electron 53(5):1398–1409 on power electronics—ECCE, Asia, pp 1110–1117
16. Kazmierkowski M, Krishnan R, Blaabjerg F (2002) Control in 21. Dasgupta S, Mohan SN, Sahoo SK, Panda SK (2013) Lyapunov
power electronics-selected problems. Academic, New York function-based current controller to control active and reactive
17. Twining E, Holmes DG (2003) Grid current regulation of a three- power flow from a renewable energy source to a generalized three-
phase voltage source inverter with an LCL input filter. IEEE Trans phase microgrid system. IEEE Trans Ind Elects 60(1):799–813
Power Electron 18(3):888–895
18. Ghartemani MK, Iravani M (2004) A method for synchronization
of power electronic converters in polluted and variable-frequency
environments. IEEE Trans Power System 19(3):1263–1270

123