7th IEEE International Symposium on Applied Computational Intelligence and Informatics May 24-26, 2012 Timioara, Romania

bIHH!IOH POUc UcVcOQcU OI bH-bCc

N bS!cH IH IS!IIDH!IOH c!WOIKS
L. hCCh-LICOCl0lu, . NIh0l-Cp0. 80I188Cn0nd H. Indn0l
POLITEHNICA University/Automation and Computer Science, Timisoara, Romania
Denmark Technical UniversitylElectrical Engineering, Roskilde, Denmark
cosmin.koch@aut.upt.ro; Imih@elektro.dtu.dk
Abstact-This paper presents a PV panel simulation model
using the single-diode four-parameter model based on data
sheet values. The model was implemented frst in
MA TLA/Simulink, and the results have been compared
with the data sheet values and characteristics of the PV
panels in standard test conditions. Moreover to point out the
strong dependency on ambient conditions and its infuence
on array operation and to validate simulation results with
measured data a complex model has also been developed. A
PV inverter model, using the same equations and
parameters as in MATLAB/Simulink has also been
developed and implemented in PowerFactory to study load
fow, steady-state voltage stability and dynamic behavior of
a distributed power system.
I. INlODUCTION
Renewable energy systems are expanding due to not
only environmental aspect but also due to social,
economical and political interest. The European Union is
aiming at a specifc CO2 reduction in the electicity sector
in the near fture (20 reduction by 2020).This wll
involve a signifcant gowth of PV installation all over
Euope resulting in a few hundred Giga watts of capacity
[I].
The increased PV capacity will infuence power system
operation and design. Power supplied fom a PV array
depends mostly on present abient conditions such as:
iradiation and temperature [2-5].
PV output voltage chages mainly with temperatue
while PV outut current changes mainly with irradiation.
Therefore in order to develop a very precise simulation
model the local wind speed and the sola radiation
incidence angle, in terms of the slope ad surface azimuth,
should be considered [5-8].
In order to determine the hourly incident radiation on a
surface of ay orientation it is necessay to evaluate the
ratio of incident radiation on the tilted suface to that on a
horizontal surface considering beam, sky difse and
gound refected raiation sepaately [3], [6], [9].
Increased distibuted generation is becoming more
importt in the curent power system ad in the fture it
will rely more on distributed energy resouces ad micro
gids. The fexible micro-grid has to be able to
importexport energy fom/to the gid, contol the active
and reactive power fows and manage of the storage
energy [7, 10, and 14].
This paper focuses on the simulation models of a small
scale PV System connected to a distibuted network ad
on improvements and validating it using experimental
facility of an active and distributed power systems
laboratory. In order to fnd out the diferences between
DER components in power systems and to study the
impact on bus voltage and fequency the system has also
been implemented in PowerFactory.
II. DISTUTED ENRGY SYSTEM ACl TCTUE.
EXERNTAL FACILITY
The experimental system is a laboratory for reseach in
distributed contol and smart grids with a high share of
renewable energy production. Its experimental facility is a
WindIV!iesel Hybrid Mini-Grid with local storage and
a novel contol infastcture. The facility is spread across
thee sites located several hundred meters apart.
It includes two wind turbines (11kW and 55kW), a PV
plant (7.8 kW), a diesel gen-set (48kW/60kVA), an
intelligent ofce building with contollable loads (up to
20kW), a nuber of loads (75 kW, 3*36 kW) and a
Vanadium Battery of 15 kW/190 kWh.
At each of the three sites there is a switchboad that
allows the components installed at the site to be connected
to either of two bus bars. The two bus bars at each site ae
connected to a crossbar switchboard allowing the fexible
setup of the system(s) to be studied. The bus bas can be
either connected to the national gid or can be pa of an
isolated system. It allows components and systems to be
in gid connected operation, island operation, or operation
in parallel with wind turbine or PV -plant.
A. PVPanels
The PV panels ae mounted in three stings: two strings
having 18 panels of 165 W each, and aother one
containing 12 panels of 100 W. The strings ae connected
to the gid through a three-phae PV inverter.
D. Data Acquiition and Control System
The data acquisition and contol system (hardwae and
sofware) is responsible for the superision and contol of
the research platform for distributed intelligent energy
systems with a high penetration of renewable energy. The
supervisory sofware code was written in Java and is able
to manage the data acquisition, processes the data and
executes the contol loop and oututs the control
variables. The sensors oututs are connected to a signal
conditioning board, which in H is connected to the data
acquisition (DAQ) boad based on a PC (SCADA
System).
978-1-4673-1014-7/12/$31.00 2012 IEEE -341 -
L. Koch- Ciobotaru et al. Simulation Model developed for a Small- Scale |VSystem in Distribution Networks
III. PV PAEL AAYMODELIG
A. Modeling of the PV Panel
This paper uses a single diode equivalent circuit for the
PV model, described by a simple exponential fnction:
' = Ise
-
__.-.
-
I

(1)
In which Ise and 10 ae the short-circuit and open-circuit
curents, R is the cell series resistance, is the number of
cells in the panel connected in series and V _ represents the
junction thermal voltage which includes the diode quality
factor, the Boltzman's constant, the temperatue at STC
and the chage of the electron.
Maufacturers tpically provide limited operational
data for photovoltaic panels. These data ae available only
at stdad rating conditions, for which the irradiance Go
is 1000 W/m
2
and the cell temperature Teel is 25 C
except for the NOCT which is determined at 800 W/m
2
ad an abient temperate To of 20 C.
Equations for the short circuit curent Ise and the open
circuit voltage Vo as a fction of absolute temperatue
A T include temperate coefcients that provide the rate
of change with respect to temperate of the PV
performance paameters, can be express as:
Ise
= Isc2s . (I + /,
.
T)
Voc = Voc2S (I + Z' T) (2)
T = ell
-
T
o
To complete the model it is also necessay to tae into
accout the variation of the paaeters with respect to
irradiance:
I
sc=Isc2S(Ga/lOOO) (3)
Using a fou paameters model of a single diode
equivalent circuit, the v-i characteristics for a solar panel
string depending on irradiace ad temperatue has the
following expressions:
v
=
nps 'e +nps nsV .1n(l
-
i/(nsp .
Ise2S .Go/I
OO
)
(4)
(5)
The equations (4) and (5) can be used to calculate the
voltage and curent over a string of panels.
The model used to obtain the static characteristics of the
PV panels has been developed in M TLAB using the
equations presented above. The model was developed for
one pael, as a fction of irradiance and temperature.
The model has as inputs Go ad Tee/ on the panel and it
sweeps the voltage range of the PV panel in order to
calculate the outut curent and power. PV cells have
nonlinea i-v and p-v chaacteristics. Its output voltage
ad power chage according to temperatue and
irradiation.
Fig. 1 shows the typical chaacteristics for a PV model
and also a compaison between PV technical
chaacteristics fom datasheet (on the lef) ad simulation
results for one panel.
.meu_c_tP
4
m
tx
t
n
, Iw
'"7
>
I :
|_
'm

t . . m
^^
|mwr>q6|@
a)
b)
!Y
'
I
I

t
Y
j

c)
twm
0
Z X K K J0
WW!|
......................

'x
` %
Z w w w !w lZ
mmem
Fig. I. Comparison between PV tehnical characteristics fom
datasheet (on te lef) ad simulation results for one pael.
D. Modeling of the PV Array
A MTLAB/Simulink model has been developed,
based on te equations presented in the last section, in
order to reproduce the electrical behavior of the existing
PV strings at the experimental system. The model has the
temperature and the solar radiace on the panel as inputs
ad the AC power at the inverters' gid side as outut.
For obtaining the maximum power of the pael strings,
the condition (dp/dv=O) should be flflled.
The block diaga implemented in Simulink that was
developed to implement this model is depicted in Fig. 2.
m w I
f + =,! W6c1+
'r
Fig. 2. PV string model implemented in Simulin
-34 2 -
7th IEEE International Symposium on Applied (omputationallntelligence and Informatics May 24-26, 2012 Timioara, Romania
C Improvements of the Model
Two types of measurements are taken fom the
experimental facility: ambient measurements fom the
weather station and electical measurements taken fom
the inverter.
The three ambient measurements: abient temperatue,
horizontal sola radiation, and wind speed are fed to a
module that calculates the cell temperatue of the PV
panels and the sola radiation on them.
1) Cell Temperature
The cell temperatue Tee/ can be very diferent fom the
ambient temperature Ta and it depends on the sola
irradiation G_, Ta and also on the wind speed Ws. Solar
irradiation acts on increasing Tee/ and the wnd speed has a
cooling efect and lowers Tee/
[
3
],
[
5
]
.
If the PV panels ae mounted in the regions with high
wind potential (as in ou case), the wind speed must be
considered. The forced (wind) convection is lage for high
wind speeds and the cell temperature fnction takes the
following form
[
5
]
:
},,
=
7 +m(0.32/(8.91+2 .W IO.67)Ga (6)
Where is the mounting coefcient, which depends on
the mounting conditions of te PV panels and Ws is the
wind speed measured on horizontal plane.
The wind that produces the cooling efect through
forced convection is the wind paallel to the panel surface;
that is why the transformation _,_,j= W)0.67 is used.
For a better understanding on the infuence of sola
irradiace and wind speed on the cell temperature, a
gaphical representation of these values is depicted in Fig.
3. The differences in temperature of the PV cells
according to diferent considerations ae also presented.
2) Solar irradiance on the PV panel
The sola radiace has infuence only on the curent and
implicitly on the output power of the PV panel, so the
plots contain the efect of the chage in the input sola
iradiace, according to the 'adaptation' of the horizontal
sola irradiace to the real case of the PV pael.
b
8me0I
b

4o
..,

O
O
E

L0060B0g 8
T TD
1mB 0uG]
W$
Fig. 3. Ambient measurements and their efect on the PV cell
temperature.
The sola irradiance input to the model is the horizontal
value measured fom the weather station. This traslates
into substantial diferences as ca also be seen in Fig 3.
D. Simulation and validation of the array model
The simulation model, developed before in
MTLA/Simulink for one PV panel, has been improved
and modifed for a PV aray with three stings. A block
diagam of the model can be seen in Fig. 4.
The PV array model is based on the equations (1-3).
)hese equations represent the PV chaacteristics taking
1Oto account the no. of strings in parallel and the no. of
panels in series.
In Fig. 5 are presented the simulation results versus
measurements at diferent stages of the modeling. These
Figures also shows the importance of several factors that
have to be taken in consideration, especially PV panel tilt
angle and orientation.
.
Fig. 5 b) shows the sae wave form for outut power,
with the same peak values, and the same changes in power
due to shading efect in a synchronous manner. The
simulation has a delay of around 50 minutes. This is the
efect of the PV panels' orientation, which have a 13
deviation fom the E-W axis.
Considering the tilt agle and orientation, the infuence
of sola iradiance and wind speed on the cell temperatue
the measurements and simulations ae almost identically,
as can be seen in Fig. 5 c).
Fig. 4. Blok diag of the simulation model for PV my.
sscc

.. ....... PO1- mcas
sc
:scc
:ccc

scc
U
ccc
scc
.
PO1- modcl
..
. ..

Tmc\ours)
a)
-343 -
C. Koch-Ciobotaru et al. Simulation Model developed for a Small-Scale |VSystem in Distribution Networks
sscc
scaa
:scc
:c

+scc
D
c
scc
c
,
sscc
sc
:s
:caa

scc
U
ccc
sca
c
,
s
s
... " " Pot - mess
-Pot -mOdcl

:
,
s :c ::
Timc(s)
b)
:
,
s
T|mc \us)
c)
-Pot-mess
-Po
e
t - modc|
:c z:
Fig. 5. Solar irradiace adatation steps.
IV . SIULATION MODEL OF 1PV SYSTM
DEVELOPED I POWKACTORY FOR DISTRUTION
NETWORS
Computer models of power systems are widely used by
power system utilities to study load fow, steady-state
voltage stability and dynamic and transient behavior of
power system.
DIgSILENT PowerFactory has been chosen because
provides the ability to simulate load fow, RMS
fuctuations in the same sofwae environment. It
provides a comprehensive library of models for electrical
components in the power system [14].
The dynamic model of the PV System implemented in
PowerFactory has been built with standard components
libray and is based on the same equations used for
MATLAB/Simulink model presented before.
The blocks of the PV model, DC-Link and controller
of the Static Generator are implemented in the dynamic
simulation language DSL of DIgSILENT. DSL allows
the user to implement specifc models that ae not
standard in the DIgSILENT library and thus to create
own developed blocks either as modifcations of existing
models or as completely new models. The interal
simulation language DSL has also been used to defne the
PV characteristics and to initialize the parameters and
variables of the model.
Fig. 6 a) shows a single line diagam of the laboratory
achitecture implemented in PowerFactory.
Fig. 6 b) shows the schematic strcture of the PV
System model, developed for time-domain simulations
where a DSL model is required, including Photovoltaic
Model, DC-Link Model, PLL block and Static Generator
with its Controller. The Static Generator is a easy to use
model of any kind of static (non rotating) generators. The
common characteristic of these generators is that they ae
all connected to the gid through a static converter.
Applications are PV Generators, Storage devices, wind
generators etc.
On the basic date tab of the single line diagam it is
possible to set up the number of parallel generators and
the power ratings of one PV panel.
For load fow analysis, also shown in Fig. 6 a), the
local voltage controller could be set to three diferent
modes: cos<, V and droop.
For RMS and EMT simulations the static generator
supports two different models: controlled current and
voltage source models. In our case we use a controlled
current source model which has as inputs d-q axis
reference current coming fom the controller and d-q
reference agles (cosref and sinrej fom a PLL built-in
model.
Photovoltaic Model has as inputs irradiation G and cell
temperature tempCel, obtained fom MATLAB-Simulink
model considering the tilt agle, orientation ad the
infuence of sola irradiation ad wind speed on the cell
temperature (Fig. 4), implemented as a look-up table, as
ca be seen in Fig. 7. Also the MPP of current, power and
voltage as a fnction of time for one module are show.
a)
mV8yzw:
b)
Fig. 6. a) Single line diagam of SYSLAB confguration
implemented in PowerFactO and b) schematic block diagam of the
PV system model.
-34-
7th IEEE International Symposium on Applied (omputationallntelligence and Informatics May 24-26, 2012 Timioara, Romania

m
|


|
....._....... ...._....._
|
|
|
.--..,---- .._.....



_____.____ _ __ _____ U
| |
| |
| |
, . _____ _____

m
-m
:
|
| |
| |
|
|
|
|
.----

----

.-- . .._....._
'
| | |
|
......... .......
| |
| |
| |
_____.____ __ _____ U
| |
|

s .

...... . . _.., ,,
| |
| |
| |

a
=v=w=
_m

m
e
..... ...........,. .... ,... ..,
|

| |
........ . . . ...... ..........
| |


..-..,---- . . . .. ,.....


|
_____.__ _ _____ ._____ _____ U
| |
| | |
| |
| |
e,m
~
...

.....

.....

|
|

=

_
__
> wom+v
120,0

....._.. ... ....._.. ..._..--- 1.

| |
| |
..... L_____ ___ _ __ ._____ 0.0
l |


......... .......


-------AC

| `
,

0,0
.. ..........a,_
|
|

=

_
,
>eo ,
=0
Fig. 7. Simulation results of the PV system model implemented in PowerFactor
V. CONCLUSIONS
This paper proposes a four-parameter model of a PV
panel and a PV system, implemented in
MATLAB/Simulink, using data provided by the
manufacturer with semi-empirical equations to predict the
PV chaacteristics for any condition. PV chaacteristics
are modeled according to a single diode fou parameter
equivalent circuit and PV paameters values taken fom
the manufacturer technical data.
The paper also proposes a model that relies on ambient
data fom a local weather station, like most common in a
real situation, not fom sensors mounted on the PV panels.
The model calculates the cell temperature and the sola
iradiace on the PV panels considering, aong others,
the tilt angle, the orientation of the panels, and the wnd
cooling efect. The paper shows that these factors
sigifcantly infuence the power output fom the PV
panels.
Comparison with experimental data, acquired by
SCADA system and processed by MA TLAB, ad with the
chaacteristics of the PV paels, provided by
manufacturers, has show that the model implemented in
MA TLAB/Simulin ca be an accurate tool for the
prediction of energy production.
A PV system model, using the same equations and
paaeters as in MA TLAB/Simulink to defne the PV
module and chaacteristics, has also been developed and
implemented in PowerFactory to study load fow, steady
state voltage stability ad dynamic behavior of a
distibuted power system.
A comparison between both simulation models,
implemented in MATLAB/Simulink ad PowerFactory,
has show a good similarity. That meas that this work
can be used for frther development of tools for DER
components in a distributed network.
ACKOWLEDGMNT
This work was supported in pat by the E.U. Project
Smooth PV, No. 228449/2011 and also paially supported
by the strategic grant POSDRU/88/1.5/S/50783 (2009) of
the Minist of Labor, Faily and Social Protection,
Romania, co-fnanced by the European Social Fund -
Investing in people.
REFERENCES
[I] www.energynautics.com. Enerautics GmbH, Longen,
Germany, 2010.
[2] S. Seme, G. Stumberg, and J. Vorsic, "Maximum efciency
trajectories of a two-axis sun trackig system determined
considering tracking system consumption", |LLL rds. O
ower L/ectrocs, vol. 26, no. 4, pp. 1280-1290, Apr. 2011.
[3] W. De. Soto, S. A. Klein, and W. A. Beckman, "Improvement and
validation of a moel for photovoltaic aray perforance",
LL5LI|Lk5o/ur Lerg,vol. 80, pp. 78-88, 2006.
[4] Y. Rifonneau, S. Bacha, S. Barruel and S. Ploix, "Optimal Power
Management for gid conected PV Systems with batteries", |LLL
rusucto o 5ustuu0/e Lerg, vol. 2, no. 3, pp. 309-320,
July 2011.
[5] E. Skoplaski, A. G. Boudouvis ad J. A. Polyvos, "A simple
corelation for the operating temperature of photovoltaic modules
of arbitrar mounting", LL5LI|Lk 5o/ur Lerg Auteru/s, pp.
1393-1402,2008.
[6] K. Murat, S. Mehmet, B. Yunus, D. Sedat, "Determinig optimum
tilt agles ad orientations of photovoltaic panels in Saliurfa",
|LLL rusucto o keeu0/e Lerg, vol. 29, issue 8, pp.
1265-1275,2004.
[7] H. Jiayi, J. Chuanwen, and X. Rong, "A review on distributed
energy resources and MicroGid", LL5LI|Lk keeu0/e d
5ustuu0/eLerg keves, vol. 12, pp. 2472-2483, 2008.
[8] M. G. Villalva, J. R. Gazoli, E. R. Filho, "Comprehensive
Approach to Modelling ad Simulation of Photovoltaic Arrays",
|LLL rusucto o ower L/ectrocs, vol. 24, issue: 5, pp.
1198-1208,2009.
[9] D. Y. Goswani, Pr|nc|ples c} Sclar Eng|neer|ng, (2" ed.),
Philadelphia: Taylor &Francis, 2000, p. 81-98.
-345-
C. Koch-Ciobotaru et al. Simulation Model developed for a Small-Scale |VSystem in Distribution Networks
[10] M. Jansen, K Louie, M. E. Amoli and F. Sami, "Model and
simulaion of a 75 kW PV solar aray", i Proc. 2010 IEE PES
Transmission ad Distribution Conference and Exosition, pp. I
.
[II] S. Dezso; Teodorescu R.; ad Rodriguez P., "PV panel model
based on datasheet values", i . 2007 IEE Intetional
Symposium on Industrial Electronics, pp. 2392-2396.
[12] H. Liu, L. Jin, D. Le and A. A. Chowdhur, "Ipact of high
penetration of solar photovoltaic generation on power system
small sigal stability", in Proc. 2010 POWERCON, pp. 1-7.
[13] A.D. Hansen, P. Slsen, L.H. Hansen, H. Bindner, "Models for
a Stand-Alone PV System", Rise National Laborator, Roskilde,
Rise-R-1219(EN) /SEC-R-12, Dec. 2000.
[14] DIgSIENT PowerFactory, Digsilent gbh, November 2010.
-346-