Applied Energy xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Long term performance analysis of a standalone photovoltaic system

under real conditions
Tao Ma a,b,⇑, Hongxing Yang a, Lin Lu a
a
Renewable Energy Research Group (RERG), Department of Building Services Engineering, The Hong Kong Polytechnic University, Hong Kong, China
b
Institute of Refrigeration and Cryogenics, Shanghai Jiao Tong University, Shanghai, China

h i g h l i g h t s

Two-year environmental and operating data of a standalone PV system was collected to analyze its long term performance.

The PV array operated well with AC power generation efficiency of 10% and overall system efficiency is 7.7%.

Results suggest that it would be better to develop or integrate the PV system into a microgrid.

a r t i c l e i n f o a b s t r a c t

Article history: Renewable energy based power generation system holds the most promising solution for remote area
Received 17 June 2016 power supply. In this study, the long term operation performance of a 19.8 kWp standalone solar photo-
Received in revised form 2 August 2016 voltaic (SAPV) system in a remote island is investigated. The methodology for evaluating system perfor-
Accepted 19 August 2016
mance is introduced, and then the detailed assessment results are presented, accompanying with analysis
Available online xxxx
and discussion. Results show that the photovoltaic (PV) array operated well with AC power generation
efficiency of about 10% and overall system efficiency is about 7.7%. The reference yield, array yield and
Keywords:
final yield are 4.08, 3.05, and 2.45 kW h/kWp/day, respectively, therefore the resultant performance ratio
Performance evaluation
Stand-alone photovoltaic system
is 60%, demonstrating that the performance of SAPV plant is satisfactory during the reporting period.
Remote islands However, after long term performance monitoring and analysis, the limitation of SAPV system begins
Normalized performance parameters to arise due to lack of grid management and compensations from other energy sources, suggesting that
it would be better to develop or integrate it into a microgrid system with hybrid sources if possible.
Ó 2016 Published by Elsevier Ltd.

1. Introduction
At present there are still about 1.2 billion people still living in
dark homes, without access to utility electricity [1], instead rely
on inefficient and often dangerous alternatives such as diesel gen-
erator, kerosene lamps, candles, flashlights and batteries, or even
on power supply. As illustrated in Fig. 1, about of 95% live in
sub-Saharan Africa and South and East Asia, including 5 million
people in China [2]. Currently it is a big challenge for most of the
developing countries to provide electricity through national grid.
Solar photovoltaic (PV) system, as one of most important part of
renewable energy, provides a viable option for rural electrification
and in some areas it has been proved as a cost-effective means of
rural electrification in developing countries. In this context, during
⇑ Corresponding author at: Institute of Refrigeration and Cryogenics, Shanghai
Jiao Tong University, Shanghai, China.
E-mail address: tao.ma@connect.polyu.hk (T. Ma).
the last two decades, there has been a constant increase of the
investment in the research of PV conversion of solar radiation to
produce high quality, yet cheap solar cells and other components
of the PV systems [3], and some interesting studies have been pub-
lished concerning the utilization of PV system to supply power for
isolated consumers, such as in Mediterranean area, Middle East,
Africa, India, and Western China [4–7].
It is widely acknowledged that the efficiency and power output
of PV system under natural conditions is lower than the rated val-
ues under standard test conduction (STC), because the real operat-
ing performance is substantially affected by local environmental
conditions such as ambient temperature, relative humidity, dust
storms and suspension in air, global solar radiation intensity, spec-
trum distribution, degradation and maintenance [8–10]. In this
context, it is important to monitor and evaluate PV system’s oper-
ation performance under highly uncertain weather conditions,
because it not only provides a platform to understand its energy
production, loss mechanisms, reliability and causes of system
failures, but also assists the users to fully utilize the electricity

http://dx.doi.org/10.1016/j.apenergy.2016.08.126
0306-2619/Ó 2016 Published by Elsevier Ltd.

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
2 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx

Nomenclature

Abbreviations PSMC output power of SMC inverter (W)

AC alternating current PSI input or output power of SI inverter (W)
DC direct current EPV electricity production of PV array/plant (kW h)
PF production factor EPV electricity production of PV subarray 1 and 2 (kW h)
PR performance ratio EBin energy input into the battery bank (kW h)
PV photovoltaic EBout energy output from the battery bank (kW h)
SAPV standalone photovoltaic U PVi voltage of PV subarray 1 and 2 (V)
SI Sunny Island IPVi current of PV subarray 1 and 2 (A)
SMC Sunny Mini Center Gt incident solar radiation on unit area of tilted PV panel
SOC state of charge (W/m2)
STC standard test condition A total solar cell area (m2)
UF usage factor t time period during which solar radiation exists (h)
Eactual actual reading of PV plant output (kW h)
Symbols Enominal calculated nominal PV plant output (kW h)
gPV solar energy conservation efficiency of PV array YF final yield (kW h/kWp/day)
gSMC efficiency of the Sunny Mini Centre inverter YA array yield (kW h/kWp/day)
gB roundtrip energy efficiency of the battery bank YR reference yield (kW h/kWp/day)
gsys overall conversion efficiency of the PV plant LS system loss (kW h/kWp/day)
Pi instantaneous DC output power of the PV subarray 1 LC capture loss (kW h/kWp/day)
and 2 (W)

output and to periodically ensure that the equipment response
meets design expectations [11]. From a design perspective, the
real-time long term monitoring can also highlight potential defi-
ciencies of the power supply design, thus allowing the adoption
of adequate countermeasures [12,13].
In recent years, substantial research has been conducted to
evaluate operation performance of grid-connected PV systems
around the world in terms of system performance, energy yield
and economic effectiveness [3,11,14–18]. However, little has been
done in the field of standalone photovoltaic (SAPV) systems for
remote area power supply. For example, International Standard
IEC 61724 [19] recommended procedures for the analysis of mon-
itored data to assess the overall performance of PV systems and
help to identify any malfunction at an early stage, but it did not
provide a well-adapted method for the analysis of SAPV [13]. In
the literature, a life cycle assessment of a 4.2 kWp SAPV system
was performed in south-east of Spain [20], while the focus of this
study is energy payback time and CO2 emission reduction. A stan-
dalone lighting system powered by PV in Malaysia was assessed in
[21], which employs the simulation tool PVSYST software for pre-
dicting energy output, instead of evaluating real-time operating
data. The performance of an SAPV system in Saudi Arabia was ana-
lyzed in [8], whereas only two-month data was used and the
emphasis is the effect of high temperature on PV power generation.
The study [22] suggests that remote monitoring can strongly
increase the PV system performance, in particular for those SAPV
systems installed in remote locations. However, available studies
provide scarce information on long term monitoring and evalua-
tion of SAPV system operation under real outdoor conditions, espe-
cially in the sub-tropical weather like Hong Kong.

Fig. 1. Share of population without grid access (percent of total) [2].

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx 3

This study, therefore, contributes to understanding the SAPV

system performance in real meteorological conditions through
long-term monitoring and assessment. The evaluation method,
practical data and technical considerations, recommendations pre-
sented in this study could be used by operators and customers for
predicting SAPV system performance and for a general assessment
of the potential of PV-technology. The main objective of the current
study is to investigate the implementation, feasibility and practical
implications of PV modules and onsite SAPV system performance
in a remote island of Hong Kong. Followed our previous study in
[23], two-year completed operating and environmental data, from
August 2010 to July 2012, is collected onsite and analyzed in the
present study, to judge whether the design goals have been
achieved and to have a better understanding the long-term overall
energy performance of the SAPV system.

2. Background and PV system description

Fig. 2. The PV array on the island (19.8 kWp).

The island involved in this study, is located in the southeastern
part of the Sai Kung District of Hong Kong. The island is as much as
22 km off the coast of Hong Kong. A non-profit-making organiza-
tion, Operation Dawn, has run the Drug Addiction Treatment &
Rehabilitation Centre on this island since 1976. At present, the
island has around 60 recovering drug users and employees on
the island. Operation Dawn is contemplating extending accommo-
dation provision to include about 100 residents. As is the case of
many similar remote islands, access to the utility grid is not avail-
able. The center used to rely on power generated by 3 diesel gen-
erators that would function intermittently for hours a day, owing
to extremely high cost of the whole exercise, including expensive
transportation costs, inventory carrying cost and the high price
of diesel fuel in Hong Kong.
To improve the quality of life and produce profound develop-
mental effects for this island, a local utility company, CLP Power
Hong Kong Limited, was on the verge of deciding to install sub-
marine cables and overhead lines to provide a constant electricity
supply for this island. This proposal was finally abandoned because
of various disadvantages such as high cost and ecological destruc-
tion. After an initial investigation, a far sighted decision was taken
to use available renewable energy for a local remote power supply.
Reasons given were renewal energy is environmentally friendly,
more cost-effective than grid extension, and mature enough to
provide utilitarian quality power supply.
This RE power supply scheme is divided into two stages. During
Stage 1, completed in 2010, a stand-alone 19.8 kWp PV system, as
shown in Fig. 2, was installed on the island. This is the first stan-
dalone commercial-scale RE system in Hong Kong. This system
was mainly used to test PV system’s feasibility, understand its
operating characteristics, and prepare for the system implementa-
tion in next stage.
The schematic block circuit diagram of the PV plant is illus-
trated in Fig. 3. This is not a common type stand-alone PV system
as it employs two different types of inverter: two PV inverters (also
known as grid-tied inverters) and five bi-directional inverters (one
for standby). This architecture has higher flexibility in accommo-
dating additional generating capacities and/or load in the future.
The two Sunny Mini Center PV inverters (SMC 10000TL) directly
transform the PV output DC to supply AC to the load while surplus
energy is used to charge the battery bank via the Sunny Island
inverts (SI 5048). The SI inverter can also provide AC output using
the energy stored in the battery bank when there is high demand
or when solar energy is not available. The battery tank is composed
of 96 deep cycle cells (960 A h, 2 V). 24 cells are connected in series
to provide a 48 V nominal storage voltage. The battery bank can
absorb excess energy at times of negative net loads and deliver
energy in case of positive net loads. The whole PV plant is primarily
controlled and monitored by the SMA power Set, which is com-
posed of a series of optimally tailored components producing a
powerful and efficient AC-coupled PV off-grid system to supply
power to the residents on this island, such as lighting tube lights,
running ceiling fans, air-conditioners, fridges and other electrical
appliances.
The 19.8 kWp PV system consists of 2 subarrays with 99 solar
panels from Suntech (Fig. 4): subarray 1 (10.2 kWp) with three par-
allel strings of 17 modules connected in series and linked to one PV
inverter and the other one (9.6 kWp) with configuration of 16  3
linked to the other inverter. All the PV panels are positioned at an
inclined angle of 22.5° in a fixed direction facing south. The key
specification of the module (model: STP200-18/Ub-1) is presented
in Table 1.
3. Monitoring parameters and evaluation method
3.1. Data acquisition system and monitored parameters
To monitor system operation and evaluate its performance, the
long-term environmental data (solar radiation, wind speed, and
ambient temperature) and operating data (module temperature,
current, voltage, battery bank’s state of charge (SOC), inverter
power, and electricity consumption) have been continuously
recorded by Sunny SensorBox since the commissioning of the sys-
tem in January 2010. However, after examining the data, only
about two-year documented data is complete without any inter-
rupt and faulty. Therefore, in this study, the onsite data from
August 2010 to July 2012 are analyzed. Table 2 shows an example
of the collected parameters, which are measured each five minutes
with the help of pyranometer, thermal couple, anemometer and
infrared laser thermometer respectively. The instantaneous solar
radiation data was obtained from the Hong Kong Observatory at
an adjacent island Kau Sai Chau. The monitoring and communica-
tion signals are transmitted by the RS485 communication protocol
to Sunny WebBox, enabling the data to be transferred to the com-
puter for data backup and analysis. The data acquisition system is
controlled by a measurement-and-analysis program mounted on a
computer. The PV plant is under its commissioning from January
2010 and work properly from June 2010.
3.2. Calculation and evaluation method
Performance evaluation can assist the users to fully utilize the
PV electricity production and periodically ensure that the equip-

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
4 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx

PV subarray 1 PV subarray 2
17 x 3 panels 16 x 3 panels
(10.2kWp) (9.6kWp)

SMC PV SMC PV
inverter inverter

AC load
SI bidirectional SI bidirectional SI bidirectional SI bidirectional
inverter inverter inverter inverter

+ - + - + - + -

Battery bank Battery bank Battery bank Battery bank

(24 cells) (24 cells) (24 cells) (24 cells)

Fig. 3. Schematic block circuit diagram of the PV plant.

PV subarray 1
17 × 3 panels = 10.2kWp

PV subarray 2
16 × 3 panels = 9.6kWp

Fig. 4. PV system subarray configuration.

ment response meets design expectations. The overall power yield match the solar source. This section examines the main procedures
depends on the performance of the PV array, the efficiency of the and theoretical calculations for evaluating major components’ and
load equipment and how well the battery bank, inverter and load system’s performance.

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx 5

Table 1 Based on the electricity production and solar radiation inten-

Characteristic parameters of the PV module. sity, the instantaneous solar PV module conversion efficiency is
Characteristics Parameter Values calculated by:
Electrical Maximum power output at STC 200 W E EPV1 þ EPV2
characteristics (Pmax) gPV ¼ RA  100% ¼ R  100% ð2Þ
Maximum power (Optimum 26.2 V
A  Gt dt A  Gt dt
Operating) Voltage (Vmp)
Maximum power (Optimum 7.63 A
where gPV is the instantaneous conservation efficiency, Gi is the
Operating) Current (Imp) incident solar radiation on unit area of tilted PV panel (W/m2), A
Open circuit voltage (Voc) 33.4 V is the total solar cell area (m2), and t is the time period during which
Short circuit current (Isc) 8.12 A solar radiation exists (h), EPV1 and EPV2 are electricity production of
Module efficiency 13.6%
PV subarray 1 and 2 (kW h).
Cell efficiency 15.2%
Operating temperature 40 °C to +85 °C
Maximum series fuse rating 20 A 3.2.2. The inverter’s performance
Maximum system voltage 1000 V DC The instantaneous efficiency of the Sunny Mini Centre 10000TL
Mechanical Solar cell Polycrystalline inverter is calculated by [25].
characteristics 156  156 mm
54 (6  9) PSMC
Number of cells
gSMC ¼  100% ð3Þ
Weight 16.8 kg U PV  IPV
Dimensions 1482  992  35 mm
The Sunny Island SI5048 bi-directional inverter efficiency is
Temperature Temperature coefficient of 0.47%/°C
characteristics Pmax expressed by the following two equations:
Temperature coefficient of Voc 0.34%/°C Discharging mode:
Temperature coefficient of Isc +0.045%/°C  
 PSI 
gSI ¼    100% ð4Þ
ISI  U SI 

In reference, there are three main generally recognized moni- Charging mode:
toring methods of PV systems [11]: (1) International Standard  
I  U SI 
IEC 61724 [19], which provides PV system performance monitoring gSI ¼  SI   100% ð5Þ
PSI 
guidelines for measurements data exchange and monitoring
parameters, frequency of measurements and duration of monitor-
ing; (2) Ispra Guidelines [24] that follow an analytical monitoring 3.2.3. The battery bank’s performance
method; and (3) the normalized parameters that are developed The roundtrip energy efficiency of the battery gB , which is the
by International Energy Agency (IEA). This study is a follow-up of ratio of the energy released during discharging to the energy
our previous study [23], and the same evaluation methodology is required for restoring the battery to its original storage condition,
employed based on the guideline of IEC 61724 and new added nor- can be calculated based on Eq. (6). The SOC values of battery bank
malized parameters. All data are extracted from the documented can be obtained from the documented datasheet.
data file, as described in Section 3.1. Similar method can also be
EBout
found in [16]. The methodology for parameters calculation and gB ¼  100% ð6Þ
performance analysis are described as follows: EBin
Therefore, the overall conversion efficiency of the PV plant can
be calculated as:
3.2.1. The PV panel performance versus solar radiation and PV
conversion efficiency gsys ¼ gPV gSMC gSI gB  100% ð7Þ
To determine the relationship between the PV power output
and solar radiation, the PV output currents and voltages were
3.2.4. Normalized parameters
extracted from the documented data file. The instantaneous DC
The overall system performance was analyzed for the period
output power of the PV array (W) can be calculated by:
from August 2010 to July 2012, including the normalized parame-
ters, performance ratios, production factor, energy balances and
Pi ¼ U PV1  IPV1 þ U PV2  IPV2 ð1Þ
performance indicators (yields, losses and efficiencies). The refer-
All currents IPV1 , IPV2 and voltages U PV1 , U PV2 of PV subarray 1 ence yield Y R , the theoretically energy output from the PV plant
and 2, are monitoring parameters in collected datasheet. with nominal efficiency determined under STC. The array yield

Table 2
An example of the documented datasheet with measured parameters.

Device ID SENS0500 SENS0500 SENS0500 SENS0500 SI5048ER SI5048ER SI5048ER SI5048ER SI5048ER WR10TL08 WR10TL08 WR10TL08
Parameters IntSolIrr TmpAmb TmpMdul WindVel BatSoc BatTmp BatVtg EgyCntIn EgyCntOut E-Total Ipv Upv-Ist
hh:mm W/m2 °C °C m/s % °C V kW h kW h kW h A V
... ... ... ... ... ... ... ... ... ... ... ... ...
13:00 794.14 29.66 52.91 4.89 92.56 32.14 54.13 18449.1 13236.7 11079.75 18.18 373.36
13:05 785.27 29.7 52.77 4.21 92.74 32.2 54.23 18,450 13236.7 11080.27 17.98 375.07
13:10 780.64 29.65 52.57 3.84 92.87 32.29 54.35 18450.8 13236.7 11080.82 17.92 374.47
13:15 779.56 29.66 52.5 3.18 93.03 32.33 54.47 18451.6 13236.7 11081.38 17.89 374.69
13:20 767.38 29.47 51.99 5.5 93.16 32.4 54.65 18452.5 13236.7 11081.91 17.69 374.94
13:25 767 29.36 51.51 4.99 93.33 32.45 54.93 18453.3 13236.7 11082.46 17.61 376.6
13:30 756.29 28.97 51.37 4.03 93.46 32.48 55.15 18454.1 13236.7 11082.98 17.47 375.93
... ... ... ... ... ... ... ... ... ... ... ... ...

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
6 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx
10
8
Solar radiation (kWh/m /day)
2
6
0
Jan
Fig. 5. Mean daily global solar radiation on a horizontal surface in 2011 (kW h/m2/day).
Y A represents the number of hours per day that the PV array would
need to operate at its rate capacity to contribute the same quantity
of energy to the system as that actually measured in practice. The
final yield Y F is the usable portion of the energy derived from the
entire PV system. The calculation method for normalized parame-
ters (kW h/kWp/day) and their relationship have been presented in
our previous study [23] and other similar studies [16].
The performance ratio (PR) is one of the most important evalu-
ation indicators of a PV plant’s operating performance quality.
Specifically, PR is the quotient of the actual and theoretical energy
outputs. It shows the proportion of energy which is available for
use by the load after deduction of energy losses and energy opera-
tion consumption. Such a dimensionless PR can be used to compare
PV plants at different locations all over the world. The higher the
PR, the better the system is in utilizing its potential. The PR value
can be manually calculated using the simplified formula (8) in
[26] and data from the Sunny SensorBox and WebBox:
Eactual Eactual
PR ¼ ¼ ð8Þ
Enominal Gtotal  gSTC
where Eactual is the actual reading of PV plant output in kW h, which
is the energy consumed by the load side at the end of analysis per-
iod, Enominal is the calculated nominal PV plant output, Gtotal is the
total solar radiation incident on the entire module surface, and
gSTC is the nominal efficiency of the PV module under STC. The PR
can also be determined using the normalized parameters:
Y F ðkW hÞ YF
PR ¼ ¼  100%
Y R ðkW hÞ Y F þ LC þ LS
Another indicator proposed by the IEA PVPS Task 2 and reported
by some studies [5,27] is the production factor (PF), which is
defined as the quotient of actual array yield to the theoretically
potential output of the same PV plant under nominal efficiency, i.e.
EA YA
PF ¼ R ¼
A  gSTC  Gt dt Y R
The PF provides a comprehensive review of the factors which
explain the difference between anticipated and actual outputs.
Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
Max
75 %
50 %
25 %
Min
Feb Mar Apr May Jun Jul Aug Se p Oct No v De c
Month
4. Environmental data analysis
The operating performance of PV system under real conditions
is highly dependent on the local environmental conditions such
as solar radiation, ambient temperature, wind speed, dust storms
and suspension in air. For example, the electricity output of a PV
array is directly proportional to solar radiation intensity, while it
changes with latitude at different locations, season, time of the
day and climate conditions [10]. Therefore, in this section, the col-
lected environmental data is presented and analyzed. The effect of
environmental condition on PV array power output will be investi-
gated in Section 5.
4.1. Solar radiation assessment
The average daily global solar radiation on a horizontal surface
for each month of the year is shown in Fig. 5. The average daily
total values range from 0 to 10.3 kW h/m2/day. The yearly average
value is 4.23 kW h/m2/day. The season distribution is consistent
with the typical subtropical climate in Southern Asia. The excep-
tionally large fluctuations in the months of March and November
are mainly due to unstable climate conditions during transition
from cold to warm weather and vice versa.
4.2. Ambient and PV temperature assessment
4.2.1. PV module temperature versus ambient temperature
Fig. 6 shows the hourly mean temperatures of ambient air and
PV module in June 2011. The ambient air temperature varied from
25 °C to 35 °C, while the PV module temperature fluctuated greatly
ð9Þ from 30 °C to 60 °C. The peak difference between ambient and PV
module can reach to 30 °C. High PV temperature can result in neg-
ative impact of PV power generation at the rate of 0.47% per °C
(Table 1).
4.2.2. PV module temperature versus solar radiation
The temperature of PV modules is usually determined by many
factors such as solar radiation intensity, ambient temperature,
ð10Þ wind speed, and mounting type [11]. Fig. 7 presents an example
of the relationship between solar radiation and back surface tem-
perature of PV module in June 2011. Compared with the PV module
installed on buildings roofs or vertical facades, the back surface
 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx 7

60 temperature is not very high, ranging from 30 °C to 60 °C. The

55 Module temperature major reasons for such good heat dissipation are: (1) those PV
modules are installed in an open area, with enough space of the
50 installation from the ground; (2) this island is windy and the air-
Temperature ( )

45 flow could induce convective cooling effect for the module.

In Fig. 8, the temperatures of the PV module and ambient air on
40
two typical days (15th January in winter and 25th August in sum-
35 mer) are compared. The solar radiation profiles in these two days
30 were similar except for some fluctuations on 25th August. The
finding illustrates that both sets of ambient air temperatures
25
remained stable from sunrise to sunset, at about 31 °C for the sum-
20 Ambient temperature mer day and 14 °C for the winter day. However, the temperatures
15 of the PV module fluctuated dramatically. The module temperature
on the summer day increased from 35 °C to 55 °C, much higher
10
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 than the STC value, which would decrease PV cell conversion effi-
Day of the month ciency eventually. The variation of module temperature in the win-
ter day was much more moderate, but it still exceeded 25 °C during
Fig. 6. hourly average ambient temperature and PV module temperature in June the sunshine hours from 10:00 am to 4:00 pm. The curves in Fig. 8
2011. also illustrate that the variation of module temperature clearly, on
both the summer and winter days, has a similar trend to the ambi-
ent temperature but lagging several minutes behind solar radiation
intensity, indicating that solar radiation has higher impact on the
65 panel temperature than ambient temperature.
T=0.0255G+33.371
60 R2=0.895
5. System operation data analysis
55
Module temperature ( )

Based on the monitored data and calculation procedures pro-

50 posed in Section 3, the operation performance of the PV plant
has been evaluated with respect to PV array, inverters and battery
45 bank. The relationship between system components’ performance
and climate condition has been investigated as well. Finally, the
40
energy performance of overall system was analyzed in terms of
35 normalized parameters, performance ratios, production factor,
and energy balances.
30

25
5.1. PV array and system power output

20 5.1.1. Energy performance of two PV subarrays

0 200 400 600 800 1000 1200 Based on Eq. (1), the power output and electricity production
Solar radiation (W/m2) can be obtained. The daily electricity generated by the two PV sub-
arrays in December 2011, as an example, is given in Fig. 9. The
Fig. 7. Back surface temperature of PV module versus inclined solar radiation in daily energy production of the subarray 1, with three parallel
June 2011. stings of 16 modules connected in series, is lower than that of

1000 60
Solar_ radiation_25th Aug
Solar_ radiation_15th Jan
Amb_Temp_25th Aug
Mod_Temp_25th Aug
800 Amb_Temp_15th Jan 50
Mod_Temp_15th Jan
Solar radiation (w/m )
2

Temperature (degC)

600 40

400 30

200 20

0 10
6:00 8:00 10:00 12:00 14:00 16:00 18:00
Time

Fig. 8. The module and ambient air temperature versus solar radiation on 15th January and 25th August 2011.

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
8 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx

18 Ambient morning (6am-1pm) 8

60 Module morning (6am-1pm)
16 Morning (6am-1pm) Ambient afternoon (1pm-7pm) 7 Wind speed
Afternoon (1pm-7pm) 55 Module afternoon (1pm-7pm)
14
Power output (kW)

Wind speed (m/s)

Temperature ( )
50 Module temperature
12
5
10 45
4
8 40
Ambient temperature 3
6 35
4 2
30
2 1
25
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000 6 7 8 9 10 11 12 13 14 15 16 17 18 19

(a) Solar radiation (W/m2) (b) Solar radiation (W/m2) (c) Time

Fig. 9. Electricity production of the two PV subarrays in December 2011.

25
PV array AC power efficiency (%)

20

15

10

0
Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
2010 2011 2012

Fig. 10. Daily AC power generation efficiency of PV array during the two years.

subarray 2 with a similar configuration of 17  3. The total energy 18

production in this month is 979 kW h from subarray 1 and
1082 kW h from subarray 2, the production ratio between them 16
is 0.90, slightly lower than the capacity ratio of 0.94 (=16/17),
14
demonstrating that the subarray 2 performs a little better than
PV power output (kW)

the subarray 1. 12

10
5.1.2. PV array AC power generation efficiency
In this section, the energy production of the PV array is directly P=-0.0348T+13.58
8
obtained from the datasheet of SMC inverters, which outputs AC
power. Therefore, based on Eq. (2), the efficiency of PV array is cal- 6
culated, and it is called PV array AC power generation efficiency,
including three parts: PV module conversion efficiency, energy 4
losses in SMC inverter for transforming DC to AC power and the
2
energy losses during the PV modules switched off. Daily AC power
generation efficiency of PV array over the reporting period is pre-
0
sented in Fig. 10, which illustrating that the monthly values rang- 30 35 40 45 50 55 60
ing from 6% to 12.8%, averaged at 10%. Under real outdoor
Module temperature (°C)
conditions, the low solar radiation, higher working temperature
and disconnection of PV array can result in relative lower efficiency Fig. 11. The variation of PV array power output (a), temperature (b) and wind speed
than that under STC [23]. (c) with solar radiation (13th August 2011).

5.1.3. PV power output/efficiency versus temperature

Usually PV power output has a hysteresis due to the heat accu- the morning for the same solar radiation level. Owing to good wind
mulation on the PV modules [11]. However, this PV system demon- condition in the afternoon on the island (Fig. 11c), the ambient
strates a contrary phenomenon. As shown in Fig. 11a, the PV temperature and module temperature were lower in the afternoon
system can generate more power in the afternoon than that in than those in the morning when the solar radiation exceeded

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx 9

60
Subarray 1 (16*3)
Subarray 2 (17*3)
50

Electricity (kWh/day)
40

30

20

10

0
2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
December (2011)

Fig. 12. PV array efficiency versus module temperature (from 1st to 10th May 2011).

500 W/m2, leading to a relatively higher power generation effi- 100

ciency in the afternoon.

90
5.1.4. Hourly PV module efficiency as a function of module
temperature
The relationship between instantaneous PV array efficiency and 80
Hourly SOC (%)

module temperature power is presented in Fig. 12, demonstrating

that high temperature has negative impact on its power genera-
70
tion. When the PV module temperature changes from 30 °C to
55 °C, the PV module efficiency reduced from 12.5% to 11.5%. The
slope of the fitting trendline in Fig. 12 is 0.348, which coincides 60
with the temperature coefficient of power output provided by
the manufacturer (=0.47%/°C/13.6% = 0.0346/°C).
50

5.2. Inverter performance

Based on Eqs. (3)–(5), the efficiency of the inverters are calcu-
lated. The onsite operating data demonstrates that the SMC invert-
ers and SI inverters operates effectively with higher efficiency at
97.1% and 88.3%, respectively. It was also found that sometimes
the total power of SI inverters was positive in the daytime, indicat-
ing that power from the batteries was supplied to the load, which
usually occurred on cloudy and raining days or in the mornings.
5.3. Battery bank performance
Electricity could be consumed in different patterns when power
is generated by a PV system at a remote area. Therefore it is neces-
sary to incorporate an energy storage device into the SAPV system
because difference exists between energy produced and electricity
needed. Energy storage can hold the excess electricity and release
it during high consumption. In this section, the performance of the
energy storage system, i.e. battery bank, will be studied.
The hourly averages of the batter bank state of charge (SOC)
over the two years are presented in Fig. 13, which is similar to
the profile in 2010. It shows that the daily lowest and highest
SOC value occurred at about 7:00 am and 5:00 pm, respectively.
In the morning from 7:00 am, PV array begins to generate electric-
ity and the excess is charged to battery bank, thus an increase in
SOC can be observed. The SOC growing rate is a little higher at
the beginning before the peak solar radiation at about 1:00 pm,
and afterwards the growth of SOC continues but the rate is lower
due to increase of load consumption. This growing trend ends at
about 5:00 pm when load consumption is balanced with energy
40
0 4 8 12 16 20 24
Time
Fig. 13. Battery bank SOC’s hourly variation during two years (time interval:
5 min).
supply. Thereafter, the SOC declines continuously until 7:00 am
next morning.
It should be noted that this difference between the maximum
and minimum SOC values could be reduced if a hybrid solar and
wind system were to be used with their good complementary char-
acteristics [28]. This aspect will be further studied during Stage 2
involving an optimal hybrid solar and wind system design
[2,29–32].
5.4. System energy balance analysis
After examining the performance analysis of individual compo-
nents in the SAPV plant, this section will look into the performance
of the whole system. Fig. 14 presents the monthly electricity gen-
eration and load consumption over the reporting period. It is obvi-
ous that there was some mismatching between energy supply and
demand in the system. During the two years of operation, the
19.8 kWp standalone PV system generated total electricity of
43,157 kW h, the average daily production is 59.1 kW h, ranging
from 24 kW h (January 2011) to 98 kW h (August 2011). Daily con-
sumption is between 17 (January 2011) and 84 kW h (August
2011). It is obvious that the load side consumption coincides with

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
10 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx

100 Electricity production

Load consumption

80

Energy (kWh/day)
60

40

20

0
Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
2010 2011 2012

Month

Fig. 14. Monthly average PV electricity generation and load consumption over the reporting period.

90 losses in distribution cables. The relationship between electricity

production and consumption is illustrated in Fig. 15 with a liner
80 curve of slope at 0.86. Therefore, the overall efficiency of BOS can
also be expressed by normalized parameter usage factor (UF),
Load consumption (kWh/day)

70 which is equal to slope of the fitting line at 0.86. To improve the

energy utilization rate and PV power production, one solution is
60 Slope=0.86 to train local residents for better utilization the energy from the
R2=0.996 PV array and battery bank based on local weather condition and
50 energy stored in the battery bank (i.e. SOC) [23].
To inspect the detailed electricity production, consumption and
40 battery status during different days, three typical kinds of days are
considered, i.e. sunny days, semi-cloudy days and cloudy days,
30 which are distinguished on the basis of the average daily solar irra-
diation from 6:00 to 18:00. Three examples of each case are illus-
20 trated in Fig. 16.
In the case of sunny day (Fig. 16a, 10th May 2011), the average
10 solar irradiation between 6:00 and 18:00 was about 562.11 W/m2
20 30 40 50 60 70 80 90 100 and peaked at 1080 W/m2 at 11:45. The result illustrates that the
Electricity production (kWh/day) battery bank was under discharging status until about 7:00 am
when the energy produced by PV array was higher than consump-
Fig. 15. The relationship between monthly average electricity consumption and tion, and the battery energy crosses the critical line at 11:30. With
production.
the increase in solar radiation, the growth rate of PV output power
was faster than that of consumption, and thus the energy stored in
the time distribution of production. When PV array generates high- the batteries increase. Finally, the energy production and con-
est power with the highest solar radiation resources, the load con- sumption was 114.19 kW h and 88.39 kW h. Such surplus energy
sumption reaches its peak value with high cooling load. of 25.8 kW h was charged to battery or lost in BOS components.
The difference between electricity production and consumption In case of semi cloudy day (Fig. 16b, 29th May 2011), the aver-
results from the energy losses in the balance of system (BOS) age solar irradiation was about 274.18 W/m2. The PV array gener-
including SMC inverters, SI inverters, battery bank, and conduction ated electricity about 58.98 kW h, was slightly higher than the

1200 1200 1200

Solar radiation Solar radiation Solar radiation
electrcity output 120 electrcity output
120 electrcity output
120
Load consumption Load consumption Load consumption
1000 Battery 100 1000 Battery 100 1000 Battery 100
Solar radiation (W/m )

Solar radiation (W/m )

Solar radiation (W/m )

2
2

80 80 80
Energy (kWh)

800 800 800

Energy (kWh)

Energy (kWh)

60 60 60
600 600 600
40 40 40

400 20 400 20 400 20

0 0 0
Critical point Critical point Critical point
200 200 200
-20 -20 -20

0 -40 0 -40 0 -40

0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
(a) Time (h) (b) Time (h) (c) Time (h)

Fig. 16. Electricity generation, load consumption and battery status on 10th May 2011 (a), 29th May 2011 (b) and 24th May 2011 (c).

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx 11

6 Capture Loss
System Loss
Final Yield

Normalized parameters (h/d)

4

0
Oct

Jan

May

Jun

Jul

Oct

Jan

May

Jun

Jul
Aug

Sep

Nov

Dec

Feb

Mar

Aug

Sep

Nov

Dec

Feb

Mar
Apr

Apr
2010 2011 2012

Fig. 17. Monthly average normalized parameters results during two years.

daily energy consumption by about 1.2 kW h. The battery bank 90 12

also began to be charged at 7:00 am and it approached to the crit- 11
ical point at 12:00 and in the end of this day. 80
10

Performance Indicator (%)

In the case of cloudy day (maybe rainy day) (Fig. 16c, 24th May
70 9
2011), the average solar irradiation was low at about 132.62 W/m2
and peak at the value of 372.06 W/m2 at 12:45. Energy produced 8

Efficiency (%)
60
was 25.57 kW h, which is less than the approximated daily energy 7
consumption by about 14.4 kW h. The battery discharge time was 50 6
longer than charge time, therefore the net balance of the battery 5
energy was 14.4 kW h. In this case, the energy supply by PV is 40
4
not enough but it can be supplemented by the battery bank. How-
30 3
ever, the power supply system might fail if the weather remains
continuously cloudy for many days and batteries run out. This sit- 2
20 Perfromance ratio
uation can be overcome by using energy from the other supply Production factor 1
Overall system energy efficiency
sources, such as diesel generator or wind turbine supply, which 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
has been taken into account in the system design of stage 2.
Month

5.5. Normalized performance parameters Fig. 18. Graphic summary of the system monthly performance ratio, PV module
and overall system energy efficiency in 2011.
Based on the guideline of IEC 61724 and the calculation method
in Section 3.2.4, the normalized parameters of the SAPV system are
investigated in this section. Fig. 17 presents the monthly average the Netherlands is 0.69, followed by Germany at 0.67 and Italy at
normalized parameters results during two years including final 0.43, according to the statistics from the 140 PV systems in IEA
yield, array capture losses, and system losses. The monthly final countries [33].
yield ranged between 0.84 h/d and 4.24 h/d, with high values in The PR can be seen as an indicator of system’s technical and
summer and low values in winter. The daily system losses varied design problems and can give rise to detailed inspection of the per-
from 0.35 h/d to 0.89 h/d, and the array capture losses was in the formance of the various components. Additionally, it can provide
range of 0.34–2.72 h/d. Based on Fig. 17, the annual average reference information relevant to the system design for Stage 2,
monthly array yields were 3.20 h/d, 3.08 h/d, 2.88 h/d in 2010, the intended hybrid solar and wind system. Given that sufficient
2011 and 2012, respectively, and the final yields were 2.49 h/d, storage capacity and/or a backup generator, a greater PR than that
2.49 h/d and 2.33 h/d, respectively. The results reveal that the of the existing PV-only system should be achievable. The comple-
anticipated output (theoretical reference yield) of PV array was mentary characteristics of solar and wind resources with an effi-
4.08 kW h/kWp/day, while in fact PV array only operated as equiv- cient storage system can help electricity output to match quite
alent to 3.05 h per day (array yield) at its nominal/rated power well with consumption needs and avoid system oversizing. Train-
(19.8 kWp), and among which 2.45 h was ultimately used to meet ing local residents for better utilization of the renewable energy
the load demand (final yield). Compared to other countries, the output can also eventually improve the PR.
annual final yield of the studied system is higher than the systems Also illustrated in Fig. 18, the monthly production factor (PF)
in European countries (1.8 h/d in the Netherlands, 1.8 h/d in Ger- fluctuates greatly from 29.4% to 82%. The yearly value of 66.4%
many and 2.0 h/d in Italy), while lower than those in Asia countries indicates that about one third of the anticipated incident solar
(2.7 h/d in Japan and 3.0 h/d in Israel) [33]. energy is not converted into electricity. The reasons behind that
The difference between the two normalized performance include the factors such as high cell temperature, shading, degra-
parameters, reference yield and final yield, can be measured by a dation, dust accumulation, bad matching of electricity production
specific indicator, performance ratio (PR). As shown in Fig. 18, and consumption, and other miscellaneous reasons.
the PR in 2011 is 60%, meaning that 40% of the anticipated output Overall system efficiency during a whole year of 2011 is pre-
from incident solar energy is not converted into usable energy or sented in Fig. 18, demonstrating that the curve have similar trend
not consumed by the load. In contrast, the PR in Switzerland and of variation with PR and PF. The yearly averaged efficiency is 7.7%

Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
12 T. Ma et al. / Applied Energy xxx (2016) xxx–xxx
with the monthly value in the range 2.9–9.9%. The peak efficiency
in May is contributed by the good solar resources, high PV output
and energy consumption in that month.
6. Conclusions
This study presents a detailed inspection of long term operating
performance of a standalone 19.8 kWp PV system from the point of
the technical issues under real climate conditions. Results shows
that the overall system yield depends on the performance of the
PV plant, inverter, and how well the battery bank and load match
the solar source. Based on the data monitoring and evaluation, it
can be concluded that the PV system worked effectively during
the two years, indicating the design expectations have been
achieved. The PV generator worked well with AC power generation
efficiency at 10% and overall system efficiency is about 7.7%. The
study also demonstrates that the temperature has adversely effect
on power generation, particularly in the summer months. PV array
power generation may also limited the battery bank floating charg-
ing voltage and state of charge. The results reveal that the antici-
pated output, i.e. theoretical reference yield, of PV array was
4.08 kW h/kWp/day, while in fact PV array operated as equivalent
to 3.05 h per day (array yield) at its nominal/rated power
(19.8 kWp), and among which 2.45 h was ultimately used to meet
the load demand (final yield).
The mismatching between electricity production and load con-
sumption not only results from the energy losses in the balance of
system (BOS) including SMC inverters, SI inverters, battery bank,
and conduction losses in distribution cables, but also depends on
the energy utilization pattern of the local residents on the island.
It was found that the energy from battery bank was only a little
utilized in relation to the daily energy, resulting in surplus energy
being wasted on sunny days and insufficient energy supply on
cloudy days. Therefore it indicates that the improvements of the
PV plant can be achieved through adequate management of the
energy storage system, and the residents are suggested to be
trained for better utilization the energy from the PV array and bat-
tery bank based on weather condition and battery bank status.
Based on the evaluation method and results, this study can pro-
vide references and guidelines for general assessment of the poten-
tial of PV technology, it can also help improving system design and
operation in future development. The obtained results also indicate
the limitation of the off-grid PV system, suggesting that it would be
better to develop or integrate it into a microgrid where the excess
energy generated can be supplied to other loads while any energy
shortage can be compensated by other energy supply sources con-
nected to the microgrid, to achieve a better energy utilization ratio,
which deserves further research.
Acknowledgements
The authors would appreciate the financial supports provided
by Research Institute for Sustainable Urban Development (RISUD)
(Grant No.: K-ZJK1/1-ZVED) of The Hong Kong Polytechnic Univer-
sity, and the National Natural Science Foundation of China (NSFC)
through the Grant 51506183. CLP Power Hong Kong Limited has
provided great supports and help for data collection from the
system.
References
[1] Off-Grid Solar Market Trends Report. Bloomberg New Energy Finance and
Lighting Global, an innovation of the World Bank Group in cooperation with
Global Off-Grid Lighting Association; 2016.
Please cite this article in press as: Ma T et al. Long term performance analysis of a standalone photovoltaic system under real conditions. Appl Energy
(2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.126
[2] Ma T, Yang H, Lu L. A feasibility study of a stand-alone hybrid solar–wind–
battery system for a remote island. Appl Energy 2014;121:149–58.
[3] Milosavljević DD, Pavlović TM, Piršl DS. Performance analysis of A grid-
connected solar PV plant in Niš, republic of Serbia. Renew Sustain Energy Rev
2015;44:423–35.
[4] Li C, Yu W. Techno-economic comparative analysis of off-grid hybrid
photovoltaic/diesel/battery and photovoltaic/battery power systems for a
household in Urumqi, China. J Clean Prod 2016;124:258–65.
[5] Sharma R, Tiwari GN. Technical performance evaluation of stand-alone
photovoltaic array for outdoor field conditions of New Delhi. Appl Energy
2012;92:644–52.
[6] Baurzhan S, Jenkins GP. Off-grid solar PV: is it an affordable or appropriate
solution for rural electrification in Sub-Saharan African countries? Renew
Sustain Energy Rev 2016;60:1405–18.
[7] Yan J et al. Dynamic simulation of a PV-diesel-battery hybrid plant for off grid
electricity supply. Energy Proc 2015;75:381–7.
[8] Rehman S, El-Amin I. Performance evaluation of an off-grid photovoltaic
system in Saudi Arabia. Energy 2012;46(1):451–8.
[9] Bücher K. Site dependence of the energy collection of PV modules. Sol Energy
Mater Sol Cells 1997;47(1):85–94.
[10] Gong X, Kulkarni M. Design optimization of a large scale rooftop photovoltaic
system. Sol Energy 2005;78(3):362–74.
[11] Mondol JD et al. Long term performance analysis of a grid connected
photovoltaic system in Northern Ireland. Energy Convers Manage 2006;47
(18–19):2925–47.
[12] Kim M, Hwang E. Monitoring the battery status for photovoltaic systems. J
Power Sources 1997;64(1):193–6.
[13] Torres M et al. Online monitoring system for stand-alone photovoltaic
applications—analysis of system performance from monitored data. J Sol
Energy Eng 2012;134(3). 034502-01–08.
[14] Sharma V, Chandel SS. Performance analysis of a 190 kWp grid interactive
solar photovoltaic power plant in India. Energy 2013;55:476–85.
[15] Ayompe LM et al. Measured performance of a 1.72 kW rooftop grid connected
photovoltaic system in Ireland. Energy Convers Manage 2011;52(2):816–25.
[16] Congedo PM et al. Performance measurements of monocrystalline silicon PV
modules in South-eastern Italy. Energy Convers Manage 2013;68:1–10.
[17] Ketjoy N, Sirisamphanwong C, Khaosaad N. Performance evaluation of 10 kWp
photovoltaic power generator under hot climatic condition. Energy Proc
2013;34:291–7.
[18] Humada AM et al. Performance evaluation of two PV technologies (c-Si and
CIS) for building integrated photovoltaic based on tropical climate condition: a
case study in Malaysia. Energy Build 2016;119:233–41.
[19] 61724, I.S.I. Photovoltaic system performance monitoring – guidelines for
measurement, data exchange and analysis. Geneva, Switzerland: International
Electro Technical Commission, April; 1998.
[20] García-Valverde R et al. Life cycle assessment study of a 4.2 kWp stand-alone
photovoltaic system. Sol Energy 2009;83(9):1434–45.
[21] Waewsak J et al. Stand-alone photovoltaic (SAPV) system assessment using
PVSYST software. Energy Proc 2015;79:596–603.
[22] Tina GM, Grasso AD. Remote monitoring system for stand-alone photovoltaic
power plants: the case study of a PV-powered outdoor refrigerator. Energy
Convers Manage 2014;78:862–71.
[23] Ma T, Yang H, Lu L. Performance evaluation of a stand-alone photovoltaic
system on an isolated island in Hong Kong. Appl Energy 2013;112:663–72.
[24] Guidelines for the assessment of photovoltaic plants document B: analysis and
presentation of monitoring data. Report EUR 16339 EN. Joint Research Centre,
Ispra Establishment, Commission of the European Countries; 1995.
[25] Chel A, Tiwari GN. A case study of a typical 2.32 kWP stand-alone photovoltaic
(SAPV) in composite climate of New Delhi (India). Appl Energy 2011;88
(4):1415–26.
[26] Naxin C, Schmid J, Chen A. Data and performance evaluation of a PV hybrid
system in Gambia. In: IEEE 6th international power electronics and motion
control conference, 2009. IPEMC ’09.
[27] Mayer D, Heidenreich M. Performance analysis of stand alone PV systems from
a rational use of energy point of view. In: Proceedings of 3rd world conference
on photovoltaic energy conversion, 2003.
[28] Ma T, Yang H, Lu L. Feasibility study and economic analysis of pumped hydro
storage and battery storage for a renewable energy powered island. Energy
Convers Manage 2014;79:387–97.
[29] Ma T et al. Technical feasibility study on a standalone hybrid solar-wind
system with pumped hydro storage for a remote island in Hong Kong. Renew
Energy 2014;69:7–15.
[30] Ma T et al. Optimal design of an autonomous solar–wind-pumped storage
power supply system. Appl Energy 2015;160:728–36.
[31] Ma T, Yang H, Lu L. Development of hybrid battery–supercapacitor energy
storage for remote area renewable energy systems. Appl Energy
2015;153:56–62.
[32] Ma T, Yang H, Lu L. Study on stand-alone power supply options for an isolated
community. Int J Electr Power Energy Syst 2015;65:1–11.
[33] Jahn U, et al. International energy agency TASK II database on photovoltaic
power systems: statistical and analytical evaluation of PV operational data. In:
2nd World conference on photovoltaic solar energy conversion. Vienna,
Austria; 1998.