Energy and Power E ngineering, 2013, 5, 209-214
doi:10.4236/epe.2013.54B041 Published Online July 2013 (http://www.scirp .o rg/journal/epe)
Accommodating High PV Penetration on the Distribution
System of Kinmen Island
Yuan-Kang Wu1, Shao-Hong Tsai2, Ming-Yan Zou1
1Department of Electrical Engineering, National Chung-Cheng University
2Department of Electrical Engineering, Hwa Hsia Institute of Technology
Received February, 2013
ABSTRACT
A lar ge-scale PV installation has a profound effect on the p ower quality and operation of the grid. Therefore, as instal-
lations of PV systems increase, changes to both technologies and interconnectio n standards are needed to address these
high PV pe netrat ion systems. This paper investigates the main technical i mpacts of large-scale PV generations on power
grids and associated interconnection standards. A case study b y us ing t he PSS /E so ftware in the Kinme n po wer syst em
was impleme nted to study the steady-state and transient char a c te r istics for the high penetr ation PV systems.
Keywords: PV; Penetra tion; Interconnectio n S tandard; PSS/E; Kinmen
1. Introduction
There is an increasing introduction of renewable energy
generation into po wer grids, like wind power and photo-
voltaic (PV) generations, which causes control and oper-
ation problems to the utilities since the net wor ks have not
been pre-designed to take into account these distributed
or intermittent generations. The electricity from PV is
difficult to predict and mainly dependent on weather
conditions. Furthermore, they are usually connected to
grid through power electronic converter, and this also
presents a significant difference from conventional ge-
nerators. As the penetration of the solar PV plant in-
creases, its impact on the stability and operation of the
grid needs to be well understood. Because of the inter-
mittency of the output from solar plants, these impacts
might be significant, especially when the solar irradiance
drops from 100% to 20% in a minute, causing numerous
problems in a high penetration scenario. Additionally,
since traditional feeders are commonly designed for radi-
al unid irectional power flows, it is expected that some of
the most significant impacts occur for large PV penetra-
tion levels.
The main technical impacts of large scale PV genera-
tions on power grid are: voltage, current profiles, power
quality, protection, electric losses, power factor, power
balancing, reliability, protection and operability of the
system. Solar PV generation impacts can be of steady
state or dynamic in nature. These impacts vary in severi-
ty as a function of the degree o f penetration and location
of solar PV distributed gene rat ion. Fo r inst ance, PV ge n-
eration is genera lly at its optimal arou nd moon, and ge n-
erally this is a period of light load condition; therefore,
power export through the distribution feeder and the
substation back to the system becomes possible. The as-
sociated r everse po wer flow will te nd to ra ise the volta ge
on the distribution feeder. Additionally, introducing PV
at the load side would reduce the load demand and in
turn leads to reduced losses and improved voltage pro-
files on the feeder. This is an important benefit from a
capacity planning perspective since it allows utilities to
defer capital investments.
Typical system integration problems in large PV sys-
tems include [1]: no dispatch capability of PV solar
farms without storage, ultra-fast ramping requirements
(400 – 1000 MW/min), reactive power management of
feeders are not designed with high PV production, power
quality, especially voltage fluctuations, flicker and har-
monics may be out of IEEE-519 and other standards.
Several research works [2] indicate that the impact is
lower in the case of distributed PVs compared to concen-
trated PVs with the same penetration level, and a larger
grid with slow-response generators experiences larger
variations in system.
In the USA, the department of energy (DOE) solar
energy conducted a competitive Funding Opportunity
Announcement, and six competitively selected projects
have been completed [3]. For example, several projects
used actual distribution feeders as the basis for model
development and testing high PV penetration; according
to the research results, PV inverter capabilities may be
useful in mitigating system impacts. Intelligent inverters
are expected to p rovide significant benefit to fut ure elec-

Y.-K. WU ET AL.
210
tric distribution system operations. As PV penetration
rises, new functionality from the inverter would need to
be incorporated. Advanced PV inverters should have the
following features: reactive power control, low voltage
ride through, dynamic control, and works like a tradi-
tional synchro nous generator .
Some of the projects demonstrated how the high pene-
tration of PV will affect operations of an electric distri-
butio n gr id . O the r i mpo r tant re sea r ch it e ms i n t he se DOE
projects include the evaluation of the effects of energy
storage in combination with distributed PV generation,
the benefits from high resolution data collection, the im-
pact of PV system integra tion on p ower syste m relia bility
indices, the impact of cloud movement on power quality,
the impact of feeder configuration and the PV location on
the maximum PV penetration.
This paper investigates several potential problems as-
sociated with high penetration levels of grid-tied PV.
Additionally, interconnection codes or standards for PV
generation and major requirements will be studied. Fi-
nally, a case study by using the PSS/E software in the
Kinmen power system was implemented to study the
steady-state and transient characteristics for the high pe-
netration PV systems.
2. Potential Problems Associated with High
Penetration Levels of Grid-tied PV
The International Energy Agency (IEA) has produced a
series of reports on Task V of the PV power systems
(PVPS). The demonstration projects issued by the IEA
have revealed that voltage rise is one of the primary
concerned problems [4]. The study in [5] also indicates
that PV penetration is limited to approximately 33% by
voltage rise issues in the UK. A study from Japan [6]
suggested that voltage rise is the most serious problem
and would limit penetration levels between 5% and 20%.
Any injection of power into a distribution system would
cause a voltage rise at the point of connection, and this
must be considered by network operators.
Voltage dips due to cloud transients might be an issue
at high PV penetration. Cloud cover and morning fog
require fast ramping and power balancing. Several re-
search works [7] examined cloud transient effects if the
PV were deployed as a central-station plant, and the re-
sults indicate that the maximum tolerable system level
penetration le vel of PV wo uld be limited by the tra nsient
following capabilities (ramp rates) of the conventional
generators. Additionall y, slo w volta ge reg ulatio n a nd fast
voltage regulation in high penetration PV scenarios
should also be addressed. To mitigate the effects of PV
output variability, energy storage systems are another
option, in which the PV inverter control system is inte-
grated into the system. The storage system will generate
or absorb energy as the PV power fluctuates according to
the available solar resource.
In terms of power quality, the results of several re-
search works reveal that the PV contribution to voltage
distortion is far les s than t he c ontributio ns made b y many
customer loads [8]. Therefore, harmonics would not be a
problem as long as the inverter has a well design.
3. Interconnection Codes and Standards
Grid connection guidelines are a major subject with re-
gard to PV gene rat ion. The connection rules and technic-
al requirements that differ from region to region make it
more complicated. IEC has been developing many stan-
dards re lated to individual DG te chnologie s includi ng P V
generators. These guidelines would serve as input to de-
velop a more general connecting guideline for all types
of distributed generation systems and their interactions
with the grid. In 2000, the IEEE published IEEE Stan-
dard 929 IEEE Recommended Practice for Utility Inter-
face of Photovoltaic (PV) Systems up to 10 kW [9]. Thi s
standard was updated from IEEE Standard 929-1988 by
incorporating developments in the PV industry and by
coordinating with UL1741 1999; meanwhile, the IEEE
Standards Coordinating Committee 21 on Fuels, Photo-
voltaics, Dispersed Generation, and Energy Storage
formed a working group to develop the IEEE Standard
for Interconnecting Distributed Resources with Electric
Power Systems, or IEEE Standard 1547 [10]. It provides
a uniform standard for interconnection of distributed re-
sources with electric power systems, and also provides
requirements relevant to performance, operation, testing,
safety and maintenance of the interconnection. In Europe,
every country has its own technical requirements. How-
ever, major requirements in these interconnection codes
for distributed generators or PV generation include gen-
eral requirements and specifications, safety and protec-
tion requirements, and power quality requirements.
For general requirements and specifications, most
standards use a power size to limit the scope. IEEE 929
uses 10 kW and limits the technology to PV systems,
while IEEE 1547 uses 10 MVA and attempts to cover all
technologies. In addition, different PV interconnection
standards specify different requirements on voltage vari-
ation, system frequency, synchronization, and voltage
unbalance. As fo r safety and protection requirements, PV
generators must detect and respond to abnormal condi-
tions occurring on the power distribution systems. For
instance, in almost all technical requirements and stan-
dards, unexpected islanding operation is not wanted. PV
units must be disconnected as soon as possible when the
mai n grid is not energized. All interconnected PV plants
must have a readily accessible, lockable, visible-break
isolation switch s hall be located between the DG unit and
the gri d. T hi s d isconnectio n switc h must ac cessible to the
distribution co mpany. In addition, after grid fault, the

Y.-K. WU ET AL.
211
reconnection of the DG should be automatic once the
grid voltage and frequency have returned to and main-
taine d normal r anges.
PV generators should provide electric power to the
grid with acceptable power quality. Generally, the re-
quirements of interconnection standards include the lim-
its on ha rmonic, DC current injectio n, flicker, and power
factor. PV units using power electronic converters may
inject harmonic currents in the network. The level of
harmonics produced by the PV sho uld not cause a ny dis -
turbances on the distribution system. Many interconnec-
tion codes require specific values for different order
harmonics and the total harmonic distortion. Flicker can
be the result of fast variations of power output of gene-
rators or rapid changes in load current like arc furnaces
or induction motor starting leading to significant voltage
changes on the feeder. In order to reduce or avoid flicker,
many countries require the maximum installed power is
several times smaller than the level of the short-circuit
capacity at the power common coupling (PCC) point.
The PV unit shall not create objectionable or observable
flicker for other customers on the power system. DC in-
jection is a problem because the increased DC voltage
has the potential to increase saturation of magnetic com-
ponents, such as cores of distribution transformers. This
saturation, in turn, causes increased power system distor-
tion. Following IEEE 1547, DC currents injected by DG
must be smaller tha n 0.5% of its rated current at the con-
necti on po int.
4. Dynamic Model of PV Arrays
4.1. Modeling of PV Arrays
The physics based relationship between the output cur-
rent and vol tage of the PV cell is given as [11]:
( )
1
s
qV RI
s
nKT
ph Osh
V RI
II IeR
+

+
=−− −


(1)
where, I is output current from the cell, V is output vol-
tage of the cell, Iph is the generated current from photo-
voltaic actio n, Io is the reverse saturation current, q is the
charge of an electron, k is B oltzmann’s con stant, T is the
ambient temper a ture and n is the ideality factor.
4.2. Dynamic PV Model in PSS/E
The PV model was built using the user defined model
integration feature of PSS/E. The PSS/E Solar PV Unit
dynamic stability model was developed to simulate per-
formance of a photovoltaic (PV) plant connected to the
grid via a power converter. The model is based on the
generic type 4 wind model, with the added ability to si-
mulate output changes due to solar irradiation. The PV
Generic Model comprises the following modules: power
converter/generator module, electrical control module,
linear model of a panel's output curve, and linear solar
irradiance profile.
5. Case Study
The Kinmen power system (Figure 1) is chosen as the
test system to integrate PV plants in PSS/E. Before add-
ing any PV plants to the system, the load flow and dy-
namic study have been conducted in this work. Then, a
few case studies were performed where PV plants are
integrated at different buses (1101, 1104 and 1105) and
the effect of the changes in solar irradiation and loss of
PV plants for various penetration levels was observed.
5.1. Effect of Sudden Change of Solar Irradiance
on System Stability
PV power is characterized as an intermittent source of
energy as it is largely dependent on the environmental
conditions. The solar irradiance can have 70~80% drop
in a minute. The drop out of a large PV power in such a
short time could have significant i mpact on the system.
To investigate the impact on the system due to sharp
changes in the level of radiation, the solar irradiance has
been reduced from 1000 W/m2 to 0 W /m2 using different
ramping down rate. The objective is to find out the criti-
cal rate beyond which the system might become unstable
or require some load shedding action. Figure 2 shows
the rate at which output power was changed due to the
drop in solar irradiance at different rate, where PV pene-
tration is 30%. While the change in power for PV is
sho wn in F ig ure 2, t he cha ng e in the s yste m fr equency i s
shown in Figure 3. In this work, various scenarios with
different ramping down rates on radiation and PV pene-
tration have been simulated and the results are shown in
Table 1. It was found that in 30% PV penetration, the
frequency of the PV connected bus for the case of solar
Figure 1 . One-line diagram of the Kin men power s ystem.

Y.-K. WU ET AL.
212
Figure 2 . PV out put pow er with 3 0 % PV pe netr atio n under
diff erent ramping down rate.
Figure 3. Frequency response with 30% PV penetration in
the Kinmen syst em.
Table 1. Frequency response under various scenarios.
Penetration 5 s 10 s 30 s 60 s
5% 59.68 Hz 59.81 Hz 59.93 Hz 59.95 Hz
10% 59. 35 Hz 59.62 Hz 59.85 Hz 59.89 Hz
20% 58. 73 Hz 59.26 Hz 59.71 Hz 59.79 Hz
30% 58. 10 Hz 58.89 Hz 59.57 Hz 59.68 Hz
irradiation change (from 1000 W/m2 to 0 W/m2) in 5s
drops to 58.1 Hz which is above the limit for a
low-frequency load-shedding rela y in the Kinmen power
system to operate. If the PV penetration is continuously
increased, we can find the critical PV penetration at
which the frequency response indicated that the
load-frequency load-shedding relay started to operate.
The critical PV penetration in this study case under the
solar irradiance drop of 1000 W/m2 in 5 s is 42.5%.
5.2. Effect of Tripping all of PV Generation
In this case, the expected worst case scenarios are dem-
onstrated by assuming PV generator tripping. The PV
penetration under 5%, 10%, 20% and 30% is simulated
respectively. It is assumed that all of the PV generation
are tripped out at time t = 2 s. Fig ure 4 and Figure 5,
show the system frequency and the voltage at bus 2201
respectively when all of the PV generators are tripped.
Fro m Figures 4 and 5, it can be observed that penetration
level of PV power has significant impact on the system
frequency and voltage. With the increase of penetration
level, the severity of impact on system stability also in-
creases. A trip of the PV generation creates momentary
oscillations in the frequency. In Figure 4, the magnitude
of the frequency perturbation is influenced by the pene-
trat ion leve l. The higher the pene tratio n, the higher is the
magnitude of oscillation. For instance, for 30% penetra-
tion level, the system frequency goes down to 57.63 Hz.
Figure 4. Frequency response in the Kinmen system if PV
generati on trips of fline.
Figure 5. Transient voltage at bus 2201 in the Kinmen sys-
tem if PV generation trips offlin e.

Y.-K. WU ET AL.
213
If the PV penetration is continuously increased, we can
find the critical PV penetration at which the frequency
response indicated that the load-frequency load-shedding
relay started to operate. The critical PV penetration in
this study case (all of PV trip o ffline) is 34%.
5.3. Effect of a Three-phase Short Circuit Fault
Another worst case scenario, a three phase short circuit
fault, is applied at bus 2201 where the PV plant is con-
nected. The fault was applied through ver y small imped-
ance and lasted for 0.2 s. The solar irradiance is assumed
to remain constant at 1000 W/m2 and therefore, the out-
put po wer fro m the P V plant was at its max imum.
Due to the fault, the voltage at bus 2203 goes down to
0.058 pu until the fa ult is cle ar ed . As sho wn in Fig ure 6,
when the fault was cleared, the voltage curve shows
some transient with a maximum to a value of 1.11 p.u.
volt. However, the oscillation damped down shortly and
goes back to a normal value. Figure 7 shows the system
Figure 6. Voltage respo nse i n the K inmen s yste m if a three-
phase short circuit fa ult appe a rs.
Figure 7. Frequency response in the Kinmen system if a
three-pha se shor t ci rcuit fault appears.
frequency response. The frequency shows oscillatio n and
varies between 61.06 Hz and 59.58 Hz under 5% pene-
tration, which does not trigger a series of loads shedding
relays. If the PV penetration increases, the oscillation
amplitude reduces.
5.4. Effect of PV on Steady-state Voltage and
System Loss
Load current flowing through line impedances and dis-
tribution transformers causes voltage drops along the
feeder. Introducing distributed PV reduces the current
being drawn from the substation and improves the vol-
tage drop. The performance of grid at the point of con-
nection for PV generators is an important factor for in-
creasing the PV penetration. In this work, two types of
PV generation configurations, i.e., concentrated and dis-
tributed, ar e considered in the simulatio n stud ies. Fo r the
concentrated configuration, all of the PV generators are
installed on bus 1101. For the distributed configuration,
PV plants are integrated at different buses (1101, 1104
and 1105). Figure 8 shows the simulation results for the
voltage on bus 1101. It is found from Figure 8 that whe n
the PV penetration increases, then the voltage on bus
1101 increases. Furthermore, the type of concentrated PV
generation would cause larger voltage rise.
This work also investigates the system losses under
different PV configurations. According to the simulation
results, as 20% and 30% penetration is discussed, the
Figure 8. Voltage characteristic under different PV pene-
tration and two ty pes of PV generat ion configu r ations.
Table 2. S y st em loss under different PV penetration and
configura tions.
Penetration
(%) Distributed Type (MW) Concentrated Type (MW)
5 0.04 0.04
10 0.04 0.04
20 0.03 0.04
30 0.03 0.04

Y.-K. WU ET AL.
214
system power loss with distributed configuration type is
lower than that with concentrated one. Additionally, the
system power loss is reduced when the PV penetration is
increased under the distributed configuration. As shown
in Table 2.
6. Conclusions
As the integration of large and commercial scale PV on
the distribution level, such systems can have considera-
ble effect on the operation and protection of the system.
This work investigates the main technical impacts of
large scale PV generations on power grids and associated
interconnection standards. Additionally, a case study by
using the PSS/E software in the Kinmen power system
was implemented to study the steady-state and transient
characteristic s for t he hig h pe netratio n PV s ystems . Fro m
the simulation results, several system disturbances, such
as change of solar irradiance, tripping of PV generation,
and three-phase short circuit fault would affect the max-
imum PV penetratio n as a grid . Mor eo ver, PV generation
penetration and configuration types would affect system
voltage and power losses.
7. Acknowledgements
The authors grate fully ac kno wledge the fi nancia l support
by Bureau of Energy, Ministry of Economic Affairs,
R.O.C. under the Project "Development of Key Control
Technologies for Distributed Energy System", and the
National Science Council, Department of Executive,
R.O.C. under the grant no. NSC-101-2221-E-146-009.
REFERENCES
[1] J. H. R. Enslin and P. J. M. Heskes, “Harmonic In terac-
tion between a Large Number of Distributed Power In-
verters and the Distribution Netwo rk,” IEEE Transactions
on Power Electronics, Vol. 19, No. 6, 2004, pp.
1586-1593.
doi:10.1109/TPEL.2004.836615
[2] S. A. Pourmousavi, A. S. Cifala and M. H. Nehrir, “Im-
pact of High Penetration of PV Generation on Frequency
and Voltage in a Distribution Feeder,” North American
Power Symposium, 2012.
[3] H. Thomas, K. Lynn and A. Razon, “Current Results of
the US DOE High Penetration Solar Deployment
Project, ” 38th IEEE Photovoltaic Specialists Conference
(PVSC), 2012, pp. 731-736.
[4] A. Povlsen, “International Energy Agency Report,” IEA
PVPS T5-10: 2002, February; 2002. Available online at
www.iea.org.
[5] M. Thomson and D. Infield, “Impact o f Widespread Pho-
tovoltaics Generation on Distribution Systems,” IET
Journal of Renewable Power Generation, 007;1 (March
(1)):33–40.
[6] H. Kobayashi and M. Takasa ki, “Demonstration Study of
Autonomous Demand Area Power System,” IEEE PES
Transmission and Distribution Conference and Exhibition,
2006, pp . 548-555.
[7] R. F. Yan and T. K. Saha, “Investigation of Voltage Sta-
bility for Residential Customers Due to High Photovoltaic
Penetrations ,” IEEE Transactions on Power Systems, Vol.
27, No. 2, 2012, pp. 651-662.
[8] I. T. Papaioannou, M. C. Alexiadis, C. S. Demoulias, D.
P. Labridis and P. S. Dokopoulos, “Modeling and Field
Measurements of Photovoltaic Units Connected to LV
Grid. Study of Penetration Scenarios,” IEEE Transactions
on Pow er D elivery, Vol. 26, No. 2, 201 1, pp. 979-987.
doi:10.1109/TPWRD.2010.2095888
[9] IEEE, Standard 929-2000 IEEE Recommended Practice
for Utility Interface of Pho tovoltaic (PV) Systems, 2000.
[10] IEEE, Standard 1547, Standard for Interconnecting Dis-
tributed Resources with Electric Power Systems, June
2003.
[11] M. G. Villalva, J . R. Gazoli and E. R. Fil ho,” Co m-
prehensive Approach to Modeling andSimulation of
Photovoltaic Arrays,” IEEE Transactions on Power
Electronics, Vol. 24, No. 5, 2009, pp. 1198-1208.
doi:10.1109/TPEL.2009.2013862