Electrical Power and Energy Systems 80 (2016) 306311

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

Insertion of wind generators in electrical power systems aimed

at active losses reduction using sensitivity analysis
William M. da Rosa , Priscila Rossoni, Julio C. Teixeira, Edmarcio A. Belati
Federal University of ABC (UFABC), CECS, R. Santa Adelia, 166, 09210-170 Santo Andre, SP, Brazil

a r t i c l e

i n f o

Article history:
Received 3 April 2015
Received in revised form 24 December 2015
Accepted 4 February 2016

Keywords:
Wind energy
Sensitivity analysis
Wind generators
Incremental Transmission Losses

a b s t r a c t
The study evaluates the electrical power system behaviour when wind turbines are inserted into the
power grid. The assessment is made using a sensitivity analysis technique applied to the power flow solution. Unlike the typical algorithms, the sensitivity analysis technique does not require an iterative process, resulting in a fast method with great precision. This proposed method make easy to check the
wind turbine behaviour to the changing of wind speed. Initially, the power flow solution is obtained
and identified as the base case. When there are perturbations in the generators, the new solution is
obtained directly by sensitivity analysis technique. The technique was applied in 34-bus, 70-bus and
126-bus test distribution system. The places chosen to connect the wind turbines were determined by
the Incremental Transmission Losses method. The results demonstrate the effectiveness of the methodology. When wind turbines are inserted in the studied systems, active and reactive losses are reduced and
voltage profile is improved.
2016 Elsevier Ltd. All rights reserved.

Introduction
The development of alternative energy sources are growing in
the world. Wind energy conversion takes a key role in environmental issues, being a renewable source of natural energy, clean and
efficient. Wind turbines can be used both in connection with electrical grids as in isolated places [1].
Brazil has one of the largest wind potential on the planet and
one of the most promising markets for wind power generation.
The year 2013 was a historic year for wind energy in Brazil, being
contracted 4.7 gigawatts (GW) of energy (data EPE Energy
Research Company) [2], capable of supplying energy for about
8.5 million households. Currently, there are in operation in the
Northeast 21 wind farms and, in the region Southern 27 wind
farms. The forecast for 2015 is the integration of 106 more parks
and in 2016, more 254 parks. These data refer to the plants of
the type I, which are dispatched and scheduled by the National
Electric System Operator ONS [3]. This share growing of wind
generation in the Brazilian electric power matrix, especially in
the Northeast and South, has led the ONS in recent years the find
structural solutions and the development of models and tools to
organize this generation. Even with a good generation forecast,
the inherent wind intermittency does not guarantee an exact
amount of generated energy.
Corresponding author.
E-mail address: william.moreti@ufabc.edu.br (W.M. da Rosa).
http://dx.doi.org/10.1016/j.ijepes.2016.02.002
0142-0615/ 2016 Elsevier Ltd. All rights reserved.

Some homogeneity on the mean wind speed can be achieved in

large wind farms. A modern wind power generation unit typically
has the ability to produce between 1 and 3 MW of active power
using wind turbine with a horizontal axis. Several of these units
operate jointly in a wind farm. The generating capacity of the
new wind farms can reach more than 100 MW [4]. In these farms,
the fast power fluctuation of one turbine caused by the wind turbulence can be compensated by another one. However, even in this
case, there is a power fluctuation in terms of daily period.
The doubly fed induction generator (DFIG) is usually used for
wind generation. The great ability of the DFIG to control the power
factor permits to separate the study of the active and reactive power
flow. This work is based on the studies related to wind turbine
capacity to operate with different constant power factors, only limited by the converter power connected to the wind turbine [57].
One of the problems of the massive use of wind power is intermittency of winds. With the growing demand and the use of alternative energy sources such as the wind, new technologies to
determine the efficiency of the power system are needed. One of
these technologies is the Smart Grids, a promise to be the new
paradigm of the electrical industry. The Smart Grid will bring
advances toward the new technologies that will enable a better
electrical system management.
Because of these technological improvements in generating
electrical energy from an intermittent energy source, it is essential
to have an understanding on the alterations of system losses
caused by wind generation. These losses are represented by

307

W.M. da Rosa et al. / Electrical Power and Energy Systems 80 (2016) 306311

nonlinear functions, which are non-convex and not separable, not

allowing direct identification of the same [8].
This work presents a study of technical losses in distribution
systems caused by the wind energy, based on sensitivity analysis
(SA) of load balance equations of the system [9]. It comprises a sensitivity matrix that incorporates all the parameters of the network
to estimate the network parameters when there is intermittency in
the power injection.
The contribution of this work is to introduce a methodology to
allocate two wind farms into the power system to reduce its losses
based on the Incremental Transmission Losses (ITL) and to verify
the power system behaviour using the sensitivity analysis method.
The ITL was used with two techniques of sensitivity analysis, one
for allocating wind generators on the network and another to make
the electrical system analysis of the intermittency in the power
injections. The two techniques use as input data the Jacobian
matrix of the system. Therefore, these techniques used together
provide an efficient methodology for allocation and analysis of
wind generators.
This paper is organized in various sections: Section Power flow
equations presents the power flow equations; Section Sensitivity
analysis technique describes the SA methodology; Sections Determination of the wind turbine power presents the wind turbine
characteristics; Section Incremental Transmission Losses method
describes the ITL method used to allocate the wind generators; S
ection Methodology to integrate the SA in the ITL method presents the developed methodology to integrate the SA into ITL
method for wind generators; Section Results describes the case
studies; the simulation results, which demonstrate the effectiveness of the analyses. Finally, in Section Conclusion the concluding
remarks are presented.

The Newton Raphson (NR) method [10] was implemented to

solve the PF problem. This method brings a computational gain
in the sensitivity analysis because it uses the data from the last
iteration of the Jacobian matrix.
Sensitivity analysis technique
The high computational speed of the SA technique has great
importance in studies of electrical power system operation [11].
It helps in understanding the relationship existing between cause
and effect system parameters and can be used in real time applications. The proposed method incorporated this technique to calculate the effect of the wind intermittency in the power systems
voltage.
Two types of variables are considered in this study: operating
variables denoted by the vector u; and controlled variables
denoted by the vector x, where:
x state variables vector V; h.
esp
u active and reactive power injection vector (P esp
k ; Q k .
The active and reactive power flow equations, Eqs. (3) and (4)
can be written compactly as:

gx; u 0

5


Considering that x x is the solution to the vector control

specified u u that satisfies Eq. (5), then:

gx ; u 0

As a modification Du in u causes a change Dx in x , the expansion in Taylors Series of (6) until first order term gives us the
Eq. (7).

Power flow equations

gx Dx; u Du gx ; u Sx Dx Su Du

The active and reactive power injection are obtained from

imposing the Kirchhoffs Current Law at each power system bus
and can be calculated in polar form by Eqs. (1) and (2),
respectively.

The Sx is the system Jacobian matrix (J), the same obtained in

the NR last iteration as follows:

Pk V; h V k

V m Gkm coshkm Bkm sinhkm

X
Q k V; h V k V m Gkm sinhkm  Bkm coshkm

"

Sx J

where:
V k voltage magnitude at bus k.
V m voltage magnitude at bus m.
hkm difference in voltage phase between the k and m bus.
Gkm real element of the matrix Y BUS associated with the bus k
and m.
Bkm imaginary element of the matrix Y BUS associated with the
bus k and m.
m 2 k the set of all m bus having connection with the bus k.
The solution of the Power Flow (PF) problem is obtained from
the balance equations of the active and the reactive power given,
respectively, by Eqs. (3) and (4).
calc
DPk Pspe
k  P k V; h 0
calc
DQ k Q spe
k  Q k V; h 0

3
4

where spe superscript represents the values specified of power

injections at bus that are considered constant (constant power load
model) and calc superscript is the calculated values of power injections obtained from the vector of state variables V; h and system
parameters.

@ DP
@V
@ DQ
@V

The Su matrix is obtained by Eq. (9).

m2k

m2k

@ DP
@h
@ DQ
@h

" @ DP

Su

@P

@ DP
@Q

@ DQ
@P

@ DQ
@Q

The Su matrix results in the identity matrix, when considering

constant power load model, which is the model adopted in this
paper.
The combination of Eqs. (5) and (7) gives us:

Sx D x Su D u 0

10

Rearranging Eq. (10), one can obtain the correction Dx vector:

Dx S1
x Su Du

11

1
As Su is the identity matrix and S1
, the expression
x is equal to J
for the correction vector Dx is given by:

Dx J1 Du

12

The Eq. (12) can be written in matrix form as (13), wherein NPQ
is the number of load bus of the distribution system.

3
Dx1
"
7
6 Dx
@ DP
2
7
6
7
6
@@h
..
7
6
DQ
5
4
.
@h
Dx2NPQ 2NPQ
2

3
Du1
7
6 Du
2
7
6
7
6
..
7
6
5
.
2NPQ2NPQ 4
Du2NPQ 2NPQ
2

@ DP
@V
@ DQ
@V

#1

13

308

W.M. da Rosa et al. / Electrical Power and Energy Systems 80 (2016) 306311

In Eq. (13), the distribution system is represented by the substation and load buses. The vector u consists of the independent variables that are the active and reactive power injections into the
buses, and the vector x that represent the controlled variables,
magnitudes and phase of the buses voltage. The matrix system
(13) can be rewritten as (14).

Dh 2

6
6 Dh 3
6
6
..
6
.
6
6
6
6 DhNPQ
6
6 DV
6
2
6
6 DV
3
6
6
6
..
6
.
4

DV NPQ

7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
5
2NPQ

DPesp
2
6
6 DPesp
3
6
6
6 ..
6 .
6
6
6 DP esp
NPQ
6
1
J2NPQ2NPQ 6
6 DQ esp
2
6
6
6 DQ esp
3
6
6
6 ..
6 .
4
DQ esp
NPQ

3
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
5

where:
P Lkm active power losses in the line associated with the k and
m buses.
The sensitivity factor of the active power losses in relation to
the injection of active power of a wind generator defines the ITL
method. The incremental change of the total system losses, @PL ,
in relation to the change in net power injected in the bus k, @Pk ,
can be obtained by Eq. (18).

ITLk

14

According [15], Eq. (19) calculates the sensitivity of the active

power losses according to the changes in active and reactive power
injected into any bus:

"

@h
@PL
@V

2NPQ

xnew x Dx

15

2NPQ

1
qcp AR v 3
2

16

where P w is the power extracted by the wind turbine in [W], q is the

air density [kg/m3], cp is the power coefficient of the turbine, AR is
the cross-sectional area [m2] and v is the wind speed [m/s].
The wind turbine data used in this study are: AR = 37.5 m2;
q = 1.175 kg/m3; cp = 0.42; cut-in = 5 m/s; and cut-out = 15 m/s.
For wind speeds less than the cut-in speed, or greater than the
cut-out speed, the injected power is zero.
The nominal active power is 2 MW and the maximum reactive
power that can be supplied is 0.66 MVAr. To values between 13 m/
s by 15 m/s the active power is fixed at 2 MW and maximum reactive power supplied is 0.66 MW and for values 05 m/s, the turbine
does not provide potency. This generator works with a maximum
power factor of 0.95. The characteristic of the DFIG generator
was based on the work presented in [14].

Incremental Transmission Losses method

The ITL may be used as a sensitivity index to find the best installation point of a distributed generation unit. The determination of
the power system loss variation as a function the injected power
variation can be obtained from the solution of the AC power flow
to the system current topology.
Eq (17) presents the active power losses in a distribution line,
based on the p model.

17

@Q
@h
@Q
@V

" @PL #
@P k

@PL
@Q k

2NPQ 2NPQ

19
2NPQ

the active power injected in bus k.

@P L
is the index ITLq , represents the incremental losses cost of
@Q
k

the reactive power injected in bus k.

Performing the mathematical manipulation and replacing
matrix partial derivatives of P and Q with respect to h and V by
the Jacobian matrix J, we obtain the ITL formulation on Eq. (20).

ITLp
ITLq

The analysis of the kinetic energy conversion of wind to electricity was presented by the several authors [12,13]. The power
of the winds energy can be converted into electrical energy by:

@P
@h
@P
@V

where:
@PL
is the index ITLp , represents the incremental losses cost of
@P

Determination of the wind turbine power

PLkm Gkm V 2k V 2m  2V k V m coshkm

18

" @P #

The matrix system of Eq. (14) has at the right side the perturbation vector (injections of active and reactive power) multiplied by
the inverse of the J matrix. The correction vector of state variables
is at the left side of the equality.
Using the matrix system (14), new solutions for the state variables of the problem, xnew , can be obtained when perturbations
are performed in the load bus as follows:

Pw

@P L
@P k

1

2NPQ

" @PL #

J T 2NPQ 2NPQ

@hk

@P L
@V k

20
2NPQ

Methodology to integrate the SA in the ITL method

In the application of sensitivity analysis technique, the PF is
solved only once. As the solution is stored together with the values
of the Jacobian matrix, the solution spends little computational
resource. The steps for obtaining xnew with technique of the SA are:
(i) Enter system data.
(ii) Obtain operational point by PF, using NR method.
(iii) Identify the bus by using ITL, using the J matrix obtained in
the last iteration of the NR method.
(iv) Perturb the power injections with Du (insertion of intermittent wind power).
(v) Use (14) to compute Dx.
(vi) Use (15) to update xnew (update the power system voltage
magnitudes and angles).
(vii) If a new perturbation is desired go to item (iv).
(viii) If any xnew represents a non-adequate power quality, change
the farm wind power and the go to item (i); otherwise End.
The first part of the algorithm classify the buses to allocate the
power injection on a descending ranking as a function of the bus
ITL, as shown in Section Sensitivity analysis technique. Therefore,

Table 1
Ranking of best connect point ITLp .
System

Bus

Bus

34
70
126

10
62
13

26
63
120

309

W.M. da Rosa et al. / Electrical Power and Energy Systems 80 (2016) 306311

Fig. 1. Active power losses to 34 bus systems using the SA  PF.

Table 2
Comparison of losses with insert of wind turbine in the 34 bus system.
Power
MW

MVAr

Active

Reactive

0.225001
0.357293
0.533335
0.759378
1.041671
1.386463
1.800007
2.000000
2.000000
2.000000

0.000000
0.000000
0.175299
0.249595
0.342380
0.455708
0.591633
0.657368
0.657368
0.657368

0.222234300
0.218128891
0.215747624
0.210666042
0.205874449
0.19998365
0.192928974
0.184671158
0.180757261
0.180757261
0.180757261

0.065261310
0.064120038
0.063457403
0.062043307
0.060707752
0.059063546
0.057091011
0.054776929
0.053678079
0.053678079
0.053678079

6
7
8
9
10
11
12
13
14
15

Base case without wind turbine insertion.

Results
The proposed methodology has been applied in three distribution systems, containing 34, 70 and 126 bus, analyzing the insertion of the two wind turbines in each system. The systems data
used in the simulation can be found respectively [1618]. The
Studies were performed in MATLAB environment with a LG A520

1.01
1.00

0 m/s
10 m/s

0.99

voltage profile (p.u.)

Losses

V (m/s)

the bus with the ITL higher value, will be most appropriate for the
wind farm connection.
After selecting the locations where to install the wind turbines,
the wind power are simulated according to the characteristic wind
speed of that region and the power system voltage are calculated.
The methodology was applied in this study to identify the best
buses for installation of two wind turbines, in order to minimize
losses of active power in the distribution system.

15 m/s

0.98
0.97
0.96
0.95
0.94
0.93
0.92

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

system bus
Fig. 2. Voltage profile to 34-bus system for different wind speeds.

310

W.M. da Rosa et al. / Electrical Power and Energy Systems 80 (2016) 306311

Table 3
Comparison of losses with insert of wind turbine in the 70 bus system.
Power (MW)

Losses

Power (MW)

Losses

V (m/s)

MW

MVar

Active (MW)

Reactive (MVar)

V (m/s)

MW

MVar

Active (MW)

Reactive (MVar)

0.225001
0.357293
0.533335
0.759378
1.041671
1.386463
1.800007
2.000000
2.000000
2.000000

0.000000
0.000000
0.175299
0.249595
0.342380
0.455708
0.591633
0.657368
0.657368
0.657368

0.224704988
0.217544216
0.213420025
0.204233876
0.195917404
0.185832172
0.173967295
0.160393957
0.154087007
0.154087007
0.154087007

0.102047939
0.099009510
0.097258546
0.093357948
0.089823641
0.085534198
0.080482476
0.074695274
0.072002900
0.072002900
0.072002900

0.225001
0.357293
0.533335
0.759378
1.041671
1.386463
1.800007
2.000000
2.000000
2.000000

0.000000
0.000000
0.175299
0.249595
0.342380
0.455708
0.591633
0.657368
0.657368
0.657368

0.193826086
0.189661395
0.187246376
0.181790978
0.176833197
0.170761416
0.163525949
0.155110017
0.151142749
0.151142749
0.151142749

0.218277071
0.213585860
0.210865831
0.204720912
0.199137787
0.192301225
0.184156063
0.174684501
0.170220614
0.170220614
0.170220614

6
7
8
9
10
11
12
13
14
15
a

Table 4
Comparison of losses with insert of wind turbine in the 126 bus system.

Base case without wind turbine insertion.

6
7
8
9
10
11
12
13
14
15

TM

Intel Core i5-2410M CPU 2.30 GHz processor, 4 GB of RAM, OS

Windows 7 Home Premium 64-bit.
Initially, the two buses of the system which are more recommended for connect the turbines are determined by the ITL technique. Table 1 shows the two better buses for each distribution
system.
In the methodology second part, the system state was initially
obtained, called base case, with PF adopting an accuracy of
105 p.u. on the power balance equations.
From the base case, new solutions were estimated using the SA.
It was considered that the DFIG technology can control the power
factor independently of the injected active power. The active and
reactive power in the load buses were increased by 2%, 4%, 6%,
up to a maximum value 50% in all load buses simultaneously keeping constant power factor. For each 2% of disturbance, the technique SA was applied and the new system state was estimated.
To validate this part of the methodology, Fig. 1 shows the comparison chart between the active losses obtained by the AS technical and the PF method. It can be seen that the technique SA
responds very well to the distribution systems.
The following tests were performed with the insertion of wind
turbines in points determined by ITL technique. From these points
simulations using SA for each wind speed were performed maintaining the characteristic of the generator presented in Section D
etermination of the wind turbine power.

Base case without wind turbine insertion.

26. These are the connection points where the wind power
generators presented in Section Determination of the wind turbine
power were introduced.
For each wind scenario, an active power was calculated and
added to the system, keeping the power factor constant. Table 2
shows the power system loss reduction as a function of the wind
speed increase. It can be observed that for high winds, the power
was limited by the maximum wind farm power. The active losses
achieve 20% reduction with fast winds.
Fig. 2 shows the voltage profile for wind speeds of 5, 10 and
15 m/s. After the choice of the insertion point, the proposed
methodology permits to verify the effect of wind intermittency
in the system power quality. In this simulation, the faster the wind
the better the voltage profile.
70-bus distribution system
The buses 62 and 63 cause major losses on the base case of 70bus distribution system. The wind turbines are connected in these
buses and Table 3 shows the wind effect on the total losses. As in
the 34-bus distribution system, the wind increase reduces the system power losses. The active power losses can be reduced by a 30%
factor when the wind is fast enough.
Fig. 3 shows the voltage profile for wind speeds of 5, 10 and
15 m/s. In this system, the wind intermittency is critical on the
buses near the injection point.

34-bus distribution system

126-bus distribution system

The ITL method has determined that the buses that cause
greater losses to 34-bus distribution system are the buses 10 and

According to the results obtained with the proposed technique

applied to the 126-bus distribution system, the generators should

1.02
0 m/s

voltage profile (p.u.)

1.00

10 m/s
15 m/s

0.98
0.96
0.94
0.92
0.90
0.88
0.86

9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69

system bus
Fig. 3. Voltage profile to 70-bus system for different wind speeds.

311

W.M. da Rosa et al. / Electrical Power and Energy Systems 80 (2016) 306311

1.06
0 m/s

1.04

10 m/s

voltage profile (p.u.)

15 m/s

1.02

1.00

0.98

0.96

0.94

0.92

system bus
Fig. 4. Voltage profile to 126-bus system for different wind speeds.

be on the buses 13 and 120. Table 4 shows that all the power
injected reduces the losses in the system.
Fig. 4 shows the voltage profile for wind speeds of 5, 10 and
15 m/s. In this case, the wind speed affects almost all the power
system, improving the voltage profile.
Conclusion
The Sensitivity Analysis tool can be used to simulate the insertion of a wind farm without needing to perform each variation of
these wind turbines with a new calculation of power flow. The proposed method can speed up the power quality analyses to validate
the wind speed effect on the voltage profile and on the losses.
Tests conducted in 34-bus, 70-bus and 126-bus systems
showed that the two SA techniques can be used in power distribution systems. It can be seen that the good choice of the wind turbines insertion point reduces both the losses and the voltage
deviation when the wind turbine is generating power. However,
as the wind intermittency affect the voltage profile, a fast method
as the proposed one is necessary to verify the power quality on
line.
This technique may be used in addition to meet the consumption needs, also to improve the quality of energy that will be delivery to the consumer.
References
[1] Blaabjerg F, Ionel DM. Renewable energy devices and systems research.
Electr Pow Compon Syst 2015;43:8378.
[2] Tolmasquim M. Plano Decenal de Expanso de Energia 2023. In: MME MdMeE,
editor <http://www.epe.gov.br/Estudos/Documents/PDE2023.pdf/2014>.
[3] National Electric System Operator (ONS) <http://www.ons.org.br/home/2015>.

[4] DeMeo EA, Grant W, Milligan MR, Schuerger MJ. Wind plant integration [wind
power plants]. Power Energy Mag IEEE 2005;3:3846.
[5] Lund T, Srensen P, Eek J. Reactive power capability of a wind turbine with
doubly fed induction generator. Wind Energy 2007;10:37994.
[6] Engelhardt S, Erlich I, Feltes C, Kretschmann J, Shewarega F. Reactive power
capability of wind turbines based on doubly fed induction generators. IEEE
Trans Energy Convers 2011;26:36472.
[7] Murari ALL, Tabares HG, Vargas GAL, Belati EA, Sousa VA, Salles M, et al. Study
of transmission system with wind power control and optimal reactive power
flow. Przeglad Elektrotechniczny 2014;07.
[8] Belati EA, Nascimento CF, Dietrich AB, de Faria H. Sensitivity analysis applied
to nodal technical losses evaluation in power transmission systems. Int Trans
Electr Energy Syst 2014;24:17885.
[9] Rosa WMD, Belati EA. First-order sensitivity applied in power distribution
system. Przeglad Elektrotechniczny 2014;7.
[10] Tinney WF, Hart CE. Power flow solution by Newtons method. IEEE Trans
Power Apparatus Syst 1967;PAS-86:144960.
[11] Belati EA, da Costa GRM. Transmission loss allocation based on optimal power
flow and sensitivity analysis. Int J Electr Power Energy Syst 2008;30:2915.
[12] Slootweg JG, Polinder H, Kling WL. Dynamic modelling of a wind turbine with
doubly fed induction generator. Power Engineering Society Summer Meeting,
2001. vol. 1. p. 6449.
[13] Heier S. Wind energy conversion systems. Grid Integr Wind Energy. Wiley &
Sons, Ltd; 2014. p. 31117.
[14] Vargas GAL, Tabares HG, Murari ALL, Belati EA, Sguarezi-Filho AJ. Analysis of
transmission systems with optimal power flow and Dobly-fed induction
generator. Latin America Trans IEEE (Revista IEEE America Latina)
2015;13:2195201.
[15] Santos ECD. User of losses sensitivity coefficients (ITL) in determining the best
location for buses of a generation unit distributed in radial distribution
systems [masters degree]. Distrito Federal Brasil: Federal District Brazil;
2009.
[16] Chis M, Salama MMA, Jayaram S. Capacitor placement in distribution systems
using heuristic search strategies. Generation Transm Distrib IEE Proc
1997;144:22530.
[17] Baran ME, Wu FF. Optimal capacitor placement on radial distribution systems.
Power Del IEEE Trans 1989;4:72534.
[18] Yan L, Yulei L, Buhan Z, Chengxiong M. A Modified Newton-Raphson power
flow method considering wind power. Power and Energy Engineering
Conference (APPEEC), 2011 Asia-Pacific 2011. p. 15.