Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
com
Received 13 January 2011; received in revised form 24 March 2011; accepted 7 June 2011
Abstract
Decentralized electricity generation by renewable energy sources offer greater security of supply for consumers while respecting the
environment. But the random nature of these sources requires us to develop sizing rules and use these systems to exploit them. This paper
proposes an integrated PV/wind hybrid system optimization model, which utilizes the iterative optimization technique following the
Deficiency of Power Supply Probability (DPSP), the Relative Excess Power Generated (REPG), the Total Net Present Cost (TNPC),
the Total Annualized Cost (TAC) and Break-Even Distance Analysis (BEDA) for power reliability and system costs. The flow chart
of the hybrid optimal sizing model is also illustrated. With this merged model, the optimal size of PV/wind hybrid energy conversion
system using battery bank can be performed technically and economically according to the system reliability requirements. Additionally,
a sensitivity analysis was carried out in order to appreciate the most important parameters influencing the economic performances of the
hybrid system. A case study is conducted to analyze one hybrid project, which is designed to supply small residential household situated
in the area of the Center for Renewable Energy Development (CDER) localized in Bouzaréah, Algeria (36°480 N, 3°10 E, 345 m).
Ó 2011 Elsevier Ltd. All rights reserved.
0038-092X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2011.06.032
2408 A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420
to endorse the lowest investment with adequate and full use an important effect on performance, and that GCHP sys-
of the renewable energy systems (photovoltaic, wind, ther- tems are economically preferable to ACHP systems for
mal systems, etc.). Thus, Chel et al. (2009) presented the the purpose of space cooling.
methodology to evaluate size and cost of PV power system A graphical construction technique to follow the opti-
components. The PV array size is determined based on daily mum combination of PV array and battery bank for a
electrical load and number of sunshine hours on optimally hybrid solar–wind system has been presented by Borowy
tilted surface specific to the country. Based on life cycle cost and Salameh (1996). For a given load and a desired LPSP,
(LCC) analysis, capital cost (US$/kWP) and unit cost of elec- the number of batteries and PV modules were calculated
tricity (US$/kW h) were determined for PV systems such as based on the minimum cost of the system. The minimum
stand-alone PV (SAPV) and building integrated PV (BIPV). cost will be at the point of tangency of the curve that rep-
Effect of carbon credit on the economics of PV system resents the relationship between the number of PV modules
showed reduction in unit cost of electricity for SAPV and and the number of batteries. Another graphical technique
BIPV systems, respectively. The developed methodology has been given by Ai et al. (2003), Kaabeche et al. (2006)
was illustrated using actual case study on 2.32 kWP PV sys- and Markvart (1996), to optimally design a hybrid solar–
tem located in New Delhi (India). wind power generation system.
Another study on performance evaluation of 2.32 kWP Tina et al. (2006) presented a probabilistic approach
photovoltaic (PV) power system located in New Delhi based on the convolution technique (Karaki et al., 1999)
(India) has been performed by Chel and Tiwari (2010). to incorporate the fluctuating nature of the resources and
The considered PV system feeds an electric air blower of the load, thus eliminating the need for time-series data,
an earth to air heat exchanger (EAHE) used for heating/ to assess the long-term performance of a hybrid solar–wind
cooling of adobe house, computer, submersible water system for both stand-alone and grid-connected applica-
pump, etc. The outdoor efficiencies, power generated and tions. Yang et al. (2008) proposed one optimum sizing
lost in PV system components were determined using method based on Genetic Algorithms by using the Typical
hourly experimental measured data for 1 year on typical Meteorological Year data. This optimization model is pro-
clear day in each month. The energy conservation, mitiga- posed to calculate the system optimum configuration which
tion of CO2 emission and carbon credit potential of the can achieve the desired LPSP with minimum Annualized
existing PV integrated EAHE system is presented too. Cost of System. Another heuristic technique based on the
Also, the energy payback time (EPBT) and unit cost of evolutionary algorithms have been performed by Ekren
electricity were determined for both stand-alone PV and Ekren (2010) for optimizing size of a PV/wind inte-
(SAPV) and building roof integrated PV (BIPV) systems. grated hybrid energy system with battery storage. In the
One more experimental outdoor performance of a study, the objective function is the minimization of the
2.32 kWP stand-alone photovoltaic (SAPV) system for four hybrid energy system total cost.
weather types have been performed by Chel and Tiwari Bernal-Agustı́n et al. (2006) present a multi-objective
(2011). The number of days and daily power generated cor- optimization (NPC versus CO2 emissions) for a hybrid
responding to four weather types in each month were used solar/wind/diesel system with battery storage based on
to determine monthly and subsequently annual power gen- Multi-Objective Evolutionary Algorithms (MOEAs). A tri-
eration from the existing SAPV system. There are three ple multi-objective optimization to minimize simulta-
daily load profiles with and without earth to air heat neously the total cost throughout the useful life of the
exchanger suitable for three seasons like summer, winter installation, pollutant emissions (CO2) and unmet load
and rainy. The life cycle cost (LCC) analysis for the exist- has been presented by Dufo-López and Bernal-Agustı́n
ing typical SAPV system is carried out to determine unit (2008). For this task, a MOEAs and a Genetic Algorithm
cost of electricity. The effect of annual degradation rate have been used in order to find the best combination of
of PV system efficiency is also presented. components and control strategies for the hybrid system.
A performance experiments and economic analysis of a According to the methods proposed by Chedid and Rah-
horizontal ground source heat pump (GSHP) system have man (1997) and Yokoyama et al. (1994) the optimal sizes
been performed by Esen et al. (2006). The GSHP system of the PV and wind power sources and the batteries are
was compared to conventional heating methods (electric determined by minimizing the system total cost function
resistance, fuel oil, liquid petrol gas, coal, oil and natural using linear programming techniques. The total system
gas) in the economical analysis using an annualized life cost consists of both the initial cost and yearly operation
cycle cost method. It was shown that the GSHP system and maintenance costs.
offers economic advantages over the mentioned first five Yang et al. (2003, 2007) have proposed an iterative opti-
conventional heating methods. However, it is not an eco- mization technique following the loss of power supply
nomic alternative system to natural gas. Another techno- probability (LPSP) model for a hybrid solar–wind system.
economic comparison between a ground-coupled heat The number selection of the PV module, wind turbine and
pump (GCHP) system and an air-coupled heat pump battery ensures the load demand according to the power
(ACHP) system has been presented by Esen et al. (2007). reliability requirement, and the system cost is minimized.
The test results indicate that system parameters can have Another iterative optimization technique to optimize the
A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420 2409
capacity sizes of different components of hybrid photovol- mix can be evaluated to meet the load demand. The meth-
taic/wind power generation system using a battery bank odology for modeling hybrid PV/wind system components
has been presented by Kaabeche et al. (2010). The recom- is described below.
mended methodology takes into account the submodels
of the hybrid system, the Deficiency of Power Supply Prob- 3.1. Photovoltaic system model
ability (DPSP) and the Levelised Unit Electricity Cost
(LUEC). With the recommended iterative optimization The input energy to PV system is solar radiation and
method, the sizing optimization of grid-independent hybrid global solar irradiance, GG,tilt, received on a tilted surface
PV/wind power generation system can be accomplished can be evaluated from the classic equation (Posadillo and
technically and economically according to the system reli- López Luque, 2009):
ability requirements. Similarly, an iterative optimization GG;tilt ¼ Gdir;tilt þ Gdiff ;tilt þ Grefl;tilt ð1Þ
technique for a stand-alone hybrid photovoltaic/wind sys-
tem (HPWS) with battery storage is presented by Diaf where Gdir,tilt, Gdiff,tilt and Grefl,tilt are the hourly direct, sky-
et al. (2008a). The main objective of the presented study diffuse and ground reflected irradiance on the tilted surface.
is to find the optimum size of system, able to fulfill the These components are calculated from hourly global solar
energy requirements of a given load distribution, for three irradiation GG,hor on a horizontal plane. The diffuse irradi-
sites located at Corsica island and to analyze the impact of ance on a tilted surface, Gdiff,tilt, is given by Klucher (1979):
different parameters on the system size. 1 þ cos b b
In this paper, an integrated photovoltaic/wind hybrid Gdiff ;tilt ¼ Gdiff ;hor 1 þ F sin3
2 2
system optimization model, which utilizes the iterative
½1 þ F cos2 h cos3 hZ ð2Þ
optimization technique following the Deficiency of Power
Supply Probability (DPSP), the Relative Excess Power where Gdiff,hor is the horizontal diffuse solar radiation; b is
Generated (REPG), the Total Net Present Cost (TNPC), the slop angle of the PV module; h is solar incidence angle
the Total Annualized Cost (TAC) and Break-Even Dis- on a tilted surface; hz is the zenith angle and F is the mod-
tance Analysis (BEDA) for power reliability and system ulating function given by:
costs, has been presented. The developed model permits 2
F ¼ 1 ðM d Þ ð3Þ
also, the calculation of the excess energy. So, the surplus
energy produced could be used in the production of hydro- With Md the diffuse fraction defined by the following
gen from an electrolyzer for long-term energy storage, relation:
helping to improve the total efficiency of the hybrid system. M d ¼ Gdiff ;hor =GG;hor ð4Þ
Additionally, a sensitivity analysis was carried out in order
to appreciate the most important parameters influencing where GG,hor is the global solar irradiation on a horizontal
the economic performances of the hybrid system. A simu- plane.
lation software code has been developed to size system The KTCORAlger model (Chikh et al., 2000) is used for
components in order to meet the load demand in the most estimating hourly diffuse irradiance on a horizontal surface
cost effective way. from hourly global irradiance GG,hor. For three different
ranges of atmospheric transmissivity, the resulting correla-
2. Integrated system description tions are given by the following expression:
8
> M d ¼ 1 0:14M t 0:037 sinðaÞ for 0 M t 0:175
Gdiff ;hor <
A schematic diagram of a stand-alone hybrid PV/wind sys- ¼ M d ¼ 1 0:43M t 0:237 sinðaÞ for 0:175 M t 0:87
GG;hor >
:
tem is shown in Fig. 1. This integrated PV/wind hybrid system M d ¼ 0:23M t 0:74 sinðaÞ for M t > 0:175
consists of PV subsystem, wind power subsystem, battery ð5Þ
bank, power conditioners, inverter, controller and other
accessory devices and cables. The PV generator and wind tur- where sin(a) is the sun elevation and Mt the hourly clear-
bine operate conjointly to meet the load demand. When pro- ness index given by Notton et al. (2010):
duction conditions (wind speed and solar radiation) are M t ¼ GG;hor =Gextr;hor ð6Þ
available, the generated power, after satisfying the load where Gextr,hor is the hourly extraterrestrial solar irradiance
demand, will be supplied to feed the battery until it is full on a horizontal surface.
charged. On the contrary, when production conditions are The direct irradiance on a tilted surface, Gdir,tilt,
weak, the battery bank will be used to cover the energy deficit is related to the direct irradiance on horizontal surface,
between production and demand until the state of charge of Gdir,hor, as follows (Mehleri et al., 2010):
battery bank decreases to its minimum level. cos h
Gdir;tilt ¼ Gdir;hor ð7Þ
cos hZ
3. Hybrid PV/wind system model
The direct irradiance on horizontal surface, Gdir,hor, can
In order to predict the hybrid system performance, indi- be calculated as follows:
vidual components need to be modeled first and then their Gdir;hor ¼ GG;hor Gdiff ;hor ð8Þ
2410 A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420
Power Power
conditioner conditioner
Controller
Inverter
/
The ground reflectance irradiance on a tilted surface, G V ðH Þ H
¼ ð13Þ
refl,tilt, is related to the global irradiance on a horizontal V ðH ref Þ H ref
surface, GG,hor, by (Iqbal, 1983):
GG;hor ð1 cos bÞ where V(H) and V(Href) are the wind speed at hub and ref-
Grefi;tilt ¼ q ð9Þ erence height H and Href and a is roughness coefficient. The
2
value of the coefficient varies from less than 0.10 for very
We have assumed the value of “q” equal to 0.2 in the cal-
flat land, water or ice to more than 0.25 for heavily forested
culation conducted.
landscapes. The one-seventh power law (0.14) is a good ref-
Hourly power output from PV system with an area APV
erence number for relatively flat surfaces such as the open
(m2), when total solar irradiance of GG,tilt (W/m2) is inci-
terrain of grasslands away from tall trees or buildings
dent on PV surface, is given by Markvar (2000):
(Kaviani et al., 2009). Fig. 2 shows typical wind turbine
P PV ¼ gPV APV GG;tilt ð10Þ characteristics. Power output of a wind energy conversion
where gPV represents the PV generator efficiency and is gi- system PW(V) can be calculated as follows (Wen et al.,
ven by Habib et al. (1999) and Kolhe et al. (2003): 2009).
8 2
gPV ¼ gr gpc ½1 lðhcell hcell;ref Þ ð11Þ >
< P R ðA þ BV þ CV Þ; V c V V R
where gr is the reference module efficiency, gpc is the power P w ðV Þ ¼ P R ; VR V VF ð14Þ
>
:
conditioning efficiency which is equal to 1 if a perfect max- 0; otherwise
imum power tracker (MPPT) is used. l is the generator effi-
ciency temperature coefficient, it is assumed to be a The constants A, B and C are found out using the fol-
constant and for silicon cells the range of l is 0.004– lowing equations:
0.006 per (°C), hcell,ref is the reference cell temperature
(°C) and hcell is the cell temperature (°C) and can be calcu-
lated as follows (Diaf et al., 2008b):
hcell ¼ ha þ ½ðNOCT 20Þ=800GG;tilt ð12Þ
Second
performence
where ha is the ambient temperature (°C) and NOCT is the region
nominal cell operating temperature (°C). gpc, l, NOCT and Rated Power
APV, are parameters that depend upon the type of module
Power Output
The Relative Excess Power Generated (REPG), Annualized Cost (TAC), and Break-Even Distance Analy-
expressed as the ratio of power excess to the sum of load sis (BEDA; where the cost of the proposed hybrid stand-
demand during the considered period is calculated by the alone system breaks even with the cost of supplying the
following equation: load with conventional grid power) are developed for sys-
, tem configurations considering the life period and replace-
XT X T
ment costs of the individual sub-systems.
REPG ¼ EPGðtÞ EL ðtÞ ð23Þ
t¼1 t¼1
4.2.1. Total Net Present Cost (TNPC)
In case (b), if the state of charge of the battery bank The Total Net Present Cost (TNPC) analysis is an eco-
decreases to its minimum level, SOCmin, the control system nomic assessment of the cost for a number of alternatives
disconnects the load and that deficit called deficiency considering all significant costs over the life period of each
power supply (DPS) at hour t, can be expressed as: alternative, adding each option’s costs for every year and
discounting them back to a common base (present worth,
DPSðtÞ ¼ EL ðtÞ ½EGen ðtÞ þ SOCðt 1Þ SOC min ginv PW). These costs can be categorized into two types: (a)
ð24Þ recurring cost, e.g. maintenance cost for PV array and wind
generator and (b) non-recurring cost, e.g. battery replace-
The Deficiency of Power Supply Probability (DPSP), for
ment cost. The conversion of recurring costs to present
a considered period T (1 year in this study), can be defined
worth is as follows (Brown and Yanuck, 1985; Christopher
as the ratio of all the (DPS(t)) values for that period to the
et al., 2003):
sum of the load demand. This can be defined as (Prasad
h ih iLp
and Natarajan, 2006): 1þe 1þe
, 1þd 1þd
1
XT X
T
PWCrec ¼ C rec h i ð26Þ
DPSP ¼ DPSðtÞ EL ðtÞ ð25Þ 1þe
1
1þd
t¼1 t¼1
A DPSP of 1 means that the load will never be satisfied and where d and e are the interest and escalation rate, respec-
the DPSP of 0 means that the load will be always satisfied. tively. Crec is the recurring cost and Lp is the system life per-
From the above-described situations, a program is devel- iod in years. The conversion of non-recurring cost to
oped in MATLAB to size the components for each config- present worth is as follows:
uration, for a particular DPSP specified by the user. The h i h iLp
1þe 1þe
flow chart of HPWES model is illustrated in Fig. 3. In this 1þd adj 1þd adj
1
program, PPV,min, PPV,max and PW,min, PW,max represent PWCnon-rec ¼ C non-rec h i ð27Þ
1þe
the lower and higher limits of the variation interval of 1þd adj
1
the PV and wind generator rated power, respectively. DPpv
and DPW represent the variation step of the PV and wind where dadj is the adjusted interest rate given as follows, P is
power, Dt the simulation step and DSD is the step of storage number of years between two successive payments for non-
days. In this study the maximum number of storage days, recurring costs and Cnon-rec is the non-recurring costs.
NSD = 3. P
The program input data set consists of hourly solar ð1 þ dÞ
Dadj ¼ 1 ð28Þ
irradiation on a tilted plane, hourly mean values of ambi- ð1 þ eÞP 1
ent temperature and wind speed, desired DPSP, load
power requirements during the year and specifications of The Total Net Present Cost (TNPC) may then be
the system devices. Using the developed program, a set expressed as follows:
of system configurations, which satisfy the system power TNPCð$Þ ¼ IC þ PWCrec þ PWCnonrec ð29Þ
reliability requirements, can be obtained. The optimal
one is subsequently predicted on the basis of the mini- where IC is the initial cost of the system components
mum cost. (including costs of civil work, installation and connec-
tions). All the parameters in the above equation are ex-
4.2. Cost analysis pressed in their present worth.
It is judicious that economic analysis should be made 4.2.2. Total Annualized Cost (TAC)
while attempting to optimize the size of the proposed The economical approach, according to the concept of
hybrid stand-alone system contributing to an acceptable the Total Annualized Cost (TAC), is developed to be one
unit price of power produced. The economic evaluation of the best indicators of economic profitability of system
of the hybrid system has been made and the cost aspects cost analysis in this study. According to the studied hybrid
have also been taken into account for optimization of the PV/wind system, the Total Annualized Cost of system
size of the system. Using the software developed, various (which is the sum of the annualized costs of each system
costs namely, Total Net Present Cost (TNPC), Total component) is found by multiplying the Total Net Present
A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420 2413
Table 1
The costs and lifetime aspect for the system components.
Escalation rate e (%) Interest rate d (%) Lifetime (year) Maintenance cost in the first year % Unit Price (US$/W) Component
4 8 25 1% of price 3.47 PV arraya
20 3% of price 3.000 Wind turbineb
4 1% of price 0.210 Battery banka
10 0% of price 0.715 Invertera
a
http://www.solarbuzz.com.
b
Mean value of the literature data.
Table 2 Table 4
Specifications of the PV module. Specifications of the single battery.
Pmax (W) Imax (A) Vmax (V) Isc (A) Voc (V) Type DOD (%) Charging Voltage Nominal Type
55 3.16 17.4 3.40 21.7 Shell SM55 efficiency (V) capacity (Ah)
50 0.9 12 100 STECO 3000
Table 3
Specifications of the wind turbine.
Tower Cut-off Rated Cut-in Rated Type 350
high (m) speed speed speed power
VF (m/s) VR (m/s) VC (m/s) (W) 300
200
100
a 1000 DPSP 0.1%
DPSP 0.3%
900
DPSP 1%
Hourly solar irradiation (Wh/m²)
800
75
Number of PV Modules
700 1 day storage
600
500 50
400
300
25
200
100
0 0
0 1000 2000 3000 4000 5000 6000 7000 8000 8760 400 800 1200 1600 2000 2400 2800 3200 3600 4000
12
25
10 DPSP 0.1%
Wind speed (m/s)
DPSP 0.3%
8 DPSP 1%
2
15
0
0 1000 2000 3000 4000 5000 6000 7000 8000 8760
Time (hours)
10
c 45 400 800 1200 1600 2000 2400 2800 3200 3600 4000
Rated Power of Wind Turbine (W)
40
35
of the battery bank.
30
25 25
DPSP 0.1%
20 DPSP 0.3%
DPSP 1%
15
Number of PV Modules
10 20 3 days storage
5
0 1000 2000 3000 4000 5000 6000 7000 8000 8760
Time (hours)
15
Fig. 5. Meteorological conditions for optimal design. (a) Solar irradiation
on horizontal plane, (b) wind speed and (c) ambient temperature.
10
Fig. 5 during the year 2003, were recorded using a properly 400 800 1200 1600 2000 2400 2800 3200 3600 4000
data-acquisition system installed at the CDER. The annual Rated Power of Wind Turbine (W)
wind energy potential for Bouzaréah at 10 m height is
Fig. 8. System configurations for different DPSP for 3 days of autonomy
187 kW h/m2 and the annual total solar radiation on the of the battery bank.
horizontal surface is 1626 kW h/m2. On a comparative
basis between the solar insolation and wind distribution
of the site, there is a great scope for generating power from are used in system unit sizing and the generation is assumed
solar for longer periods in a year. Thus, the data recorded to keep constant in each hour interval.
2416 A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420
120 10
1 day storage TNPC for 1day storage
The relationships between system reliabilities and system 2 days storage TNPC for 2 days storage 9
100 3 days storage TNPC for 3 days storage
configurations are studied. Figs. 6–8 show the results of the 8
Number of PV Modules
7
and system configurations for different days of autonomy 80
6
of the battery bank. Fig. 6 shows the relationships for a
1 day-storage battery bank. In this figure, the curves are 60 5
Cost
hyperbolic nature. Each point of them represents a couple 4
80 10
(Figs. 7 and 8), but compared to the system with 1 day-stor- 1 day storage TNPC for 1 day storage
age battery bank, the PV module and wind turbine power 70
2 days storage TNPC for 2 days storage 9
3 days storage
TNPC for 3 days storage
are more moderate. It means the hybrid system with more 8
7
demand with less supply failure. 50
6
40 5
5.3. Impact of system configurations on the TNPC, TAC and Cost
BEDA 30
4
3
20
The configurations meeting different desired DPSP 2
requirements under different battery capacities are 10
PV Modules 1
obtained from the simulation results. After the technical
0 0
criteria, the Total Net Present Cost (TNPC), Total Annu- 400 800 1200 1600 2000 2400 2800 3200 3600 4000
alized Cost (TAC) and Break-Even Distance Analysis Rated Power of Wind Turbine (W)
(BEDA) are utilized as the economic benchmarks. The sim-
Fig. 10. System configurations and Total Net Present Cost for
ulation results are demonstrated and the relationships DPSP = 1%.
between TNPC, TAC, BEDA and power reliabilities as
well as the system configurations are analyzed. In Figs. 120 9000
9–14, the curves below in solid symbols show the results 1 day storage TAC for 1 day storage
2 days storage TAC for 2 days storage 8000
of the relationship between system reliabilities or DPSP 100 3 days storage TAC for 3 days storage
values and system configurations for different days of 7000
Total annualized cost ($)
Number of PV Modules
value of TNPC, TAC and BEDA. On the other hand, a Rated Power of Wind Turbine (W)
meticulous examination into Figs. 9–14 shows that the low- Fig. 11. System configurations and Total Annualized Cost for
est TNPC, TAC and BEDA are found when the capacity of DPSP = 0.3%.
wind turbine and the number of PV modules are both mod-
erate. It is also shown that the TNPC, TAC and BEDA for desired DPSP of 0.3% and 1% for the studied case because
1 day battery storage are lower than 2 and 3 days for the batteries are much more expensive with a short lifespan.
A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420 2417
80 9000 5
1 day storage TAC for 1 day storage
2 days storage TAC for 2 days storage 8000 1 day storage
70
3 days storage
6000
50
3
5000
40
Cost 4000
2
30
3000
20
2000
1
10 PV Modules
1000
0 0 0
400 800 1200 1600 2000 2400 2800 3200 3600 4000 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Rated Power of Wind Turbine (W) Deficiency of Power Supply Probability
Fig. 12. System configurations and Total Annualized Cost for Fig. 15. REPG versus DPSP for different days of autonomy of the battery
DPSP = 1%. bank and for the optimum couples.
4
x 10
7
120 2 Discount rate
1 day storage BEDA for 1 day storage
Capital cost
2 days storage BEDA for 2 days storage 6
Break Even Distance Analysis (km)
1.5
Number of PV Modules
80 4.5
TNPC ($)
60 1
3
Distance
40
0.5 1.5
20
PV Modules
0
0 0 0.5 1 1.5 2 2.5
400 800 1200 1600 2000 2400 2800 3200 3600 4000
Parameter variation
Rated Power of Wind Turbine (W)
Fig. 16. Sensitivity of Total Net Present Cost for a hybrid PV/wind
Fig. 13. System configurations and Break Even Distance Analysis for
system.
DPSP = 0.3%.
3000 the TNPC, TAC and BEDA are sensitive to discount rates,
direct capital costs and project life.
2000 From Fig. 16, it can be seen that for a shorter discount
rates, the TNPC values would be larger and are seen to be
1000 more sensitive. For longer discount rates, the TNPC is less
sensitive. In the case of a shorter capital cost and project
0 lifetime the TNPC is less sensitive and decrease slightly.
0.5 1 1.5 2 2.5
For longer capital cost and project lifetime, the TNPC is
Parameter variation
less sensitive and increase faintly.
Fig. 17. Sensitivity of Total Annualized Cost for a hybrid PV/wind Fig. 17 shows the sensitivity of TAC to discount rates,
system. capital cost as well as project lifetime. It can be seen that
for a shorter and longer discount rates, the TAC values
would be relatively larger and are seen to be more sensitive.
1.5
Discount rate In the case of a shorter project lifetime, the TAC values
Capital cost
Project life
would be larger and are seen to be more sensitive. For a
1.2 longer project lifetime, the TAC is less sensitive and
depends on the major component replacements only. In
addition there is a low sensitivity of TAC to capital cost
BEDA (km)
0.9
for shorter and longer capital cost. Fig. 18 shows the sen-
sitivity of BEDA to discount rates, capital cost as well as
0.6
project lifetime. It can be seen that for a shorter discount
rates, the BEDA values would be larger and are seen to
0.3 be more sensitive. For longer discount rates, the BEDA
is less sensitive. In the case of a shorter capital cost and
project lifetime the BEDA is less sensitive and decrease
0
0.5 1 1.5 2 2.5 lightly. For longer capital cost and project lifetime, the
Parameter variation BEDA is less sensitive and increase faintly.
Fig. 18. Sensitivity of Break Even Distance Analysis for a hybrid PV/wind
system. 5.6. System optimal sizing result
a compromise between increasing the number of PV panels The optimal sizing results for the DPSP of 0.3% and 1%
and capacity of wind turbine and increasing the number of (power reliability requirements) for different days of auton-
batteries to meet the reliability requirements). omy of the battery bank are shown in Table 5. For DPSP
of 0.3%, TNPC is found to vary in the range of 30,040–
47,232 $, TAC vary in the range of 2814.1–4424.6 $ and
5.5. Sensitivity analysis BEDA vary in the range 0.54–0.92 km depending on stor-
age days. For DPSP of 1%, the range of TNPC is 24,405–
The effects on Total Net Present Cost (TNPC), Total 46,856 $, the range of TAC is 2286.3–4389.4 $ and the
Annualized Cost (TAC) and Break-Even Distance Analysis range of BEDA is 0.45–0.91 km. It is clear from these
(BEDA) due to the changes in the discount rate, capital results that the TNPC, TAC and BEDA for 1 day battery
costs and project life has been studied for the hybrid storage are lower than 2 and 3 days for the studied case
Table 5
Optimal sizing results for hybrid PV/wind system for DPSP = 0.3% and 1%.
DPSP = 1% DPSP = 0.3% Item
Three days storage Two days storage One day storage Three days storage Two days storage One day storage
17 18 22 18 19 37 NPV
1 1 1 1 1 1 NW
46,856 35,067 24,405 47,232 35,443 30,040 TNPC ($)
4389.4 3285.1 2286.3 4424.6 3320.2 2814.1 TAC ($)
0.91 0.67 0.45 0.92 0.68 0.54 BEDA (km)
A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420 2419
because batteries are much more expensive with a short studied case because batteries are much more expensive
lifespan. Do not forget that these results may change for with a short lifespan.
different types of PV modules, different wind turbines, dif-
ferent costs, and different types of batteries. References
Ai, B., Yang, H., Shen, H., Liao, X., 2003. Computer-aided design of PV/
6. Conclusions wind hybrid system. Renewable Energy 28, 1491–1512.
Belfkira, R., Zhang, L., Barakat, G., 2011. Optimal sizing study of hybrid
wind/PV/diesel power generation unit. Solar Energy 85, 100–110.
In this paper, an integrated photovoltaic/wind hybrid Bernal-Agustı́n, J.L., Dufo-López, R., Rivas-Ascaso, D.M., 2006. Design
system optimization model, which utilizes the iterative of isolated hybrid systems minimizing costs and pollutant emissions.
optimization technique following the Deficiency of Power Renewable Energy 31, 2227–2244.
Supply Probability (DPSP), the Relative Excess Power Borowy, B.S., Salameh, Z.M., 1996. Methodology for optimally sizing the
combination of a battery bank and PV array in a wind/PV hybrid
Generated (REPG), the Total Net Present Cost (TNPC),
system. IEEE Transactions on Energy Conversion 11, 367–373.
the Total Annualized Cost (TAC) and Break-Even Dis- Brown, J.R., Yanuck, R.R., 1985. Introduction to Life Cycle Costing.
tance Analysis (BEDA) for power reliability and system Fairmont Press, Englewood Cliffs, NJ.
costs has been presented. The recommended model consists Chedid, R., Rahman, S., 1997. Unit sizing and control of hybrid wind-
of six main parts: the submodel of the hybrid system, the solar power systems. IEEE Transactions on Energy Conversion 12,
79–85.
technical submodel developed according to the Deficiency
Chel, A., Tiwari, GN., 2010. Stand-alone photovoltaic (PV) integrated
of Power Supply Probability (DPSP) technique for system with earth to air heat exchanger (EAHE) for space heating/cooling of
reliability evaluation, the submodel witch permits the cal- adobe house in New Delhi (India). Energy Conversion and Manage-
culation of the Relative Excess Power Generated (REPG) ment 5, 1393–1409.
and the economic submodels developed based on the con- Chel, A., Tiwari, GN., 2011. A case study of a typical 232 kWP stand-
alone photovoltaic (SAPV) in composite climate of New Delhi (India).
cepts of Total Net Present Cost (TNPC), the Total Annu-
Applied Energy 88, 1415–1426.
alized Cost (TAC) and Break-Even Distance Analysis Chel, A., Tiwari, GN., Chandra, A., 2009. Simplified method of sizing and
(BEDA) which are considered as a good indicators of eco- life cycle cost assessment of building integrated photovoltaic system.
nomic profitability in the field of renewable energy. A set of Energy and Building 4, 11172–11180.
configurations meeting the desired DPSP can be obtained Chikh, M., Maafi, A., Malek, A., 2000. Etablissement d’un modèle
Mathématique pour la fraction Diffuse de l’Irradiation Solaire en
by using the DPSP submodel. The configuration with the
ALGERIE. Revue Energies Renouvelables. CHEMSS2000, 75–81.
lowest TNPC, TAC and BEDA give the optimal one. Also, Christopher et al., 2003. On the policy of photovoltaic and diesel
the developed model permits the calculation of the excess generation mix for an off-grid site: East Malaysian perspectives. Solar
energy. So, the surplus energy produced could be used in Energy 74, 453–467.
the production of hydrogen from an electrolyzer for long- Dalton, G.J., Lockington, D.A., Baldock, T.E., 2008. Feasibility analysis
of stand-alone renewable energy supply options for a large hotel.
term energy storage, helping to improve the total efficiency
Renewable Energy 33, 1475–1490.
of the hybrid system. Additionally, a sensitivity analysis Deshmukh, M.K., Deshmukh, S.S., 2008. Modeling of hybrid renewable
was carried out in order to appreciate the most important energy systems. Renewable and Sustainable Energy Reviews 12, 235–
parameters influencing the economic performances of the 249.
hybrid system. Diaf, S., Belhamelb, M., Haddadi, M., Louche, A., 2008a. Technical and
economic assessment of hybrid photovoltaic/wind system with battery
A case study is conducted to analyze one hybrid project,
storage in Corsica Island. Energy Policy 36, 743–754.
which is designed to supply small residential household Diaf, S., Notton, G., Belhamel, M., Haddadi, M., Louche, A., 2008b.
located in the area of the Center for Renewable Energy Design and techno – economical optimization for hybrid PV/wind
Development (CDER) situated in Bouzaréah, Algeria system under various meteorological conditions. Applied Energy 85,
(36°480 N, 3°10 E, 345 m). The algorithm input data set con- 968–987.
Dufo-López, R., Bernal-Agustı́n, J.L., 2008. Multi-objective design of
sists of hourly solar irradiation on the horizontal plane,
PV–wind–diesel– hydrogen–battery systems. Renewable Energy 33,
wind speed as well as ambient temperature recorded at 2559–2572.
Bouzaréah (Algeria) for the year 2003, the desired DPSP, Ekren, O., Ekren, B.Y., 2010. Size optimization of a PV/wind hybrid
load power requirements during the year and specifications energy conversion system with battery storage using simulated
of the system devices. The integrated PV/wind hybrid sys- annealing. Applied Energy 87, 592–598.
Esen, H., Inalli, M., Esen, M., 2006. Technoeconomic appraisal of a
tem is simulated by running the developed program and the
ground source heat pump system for a heating season in eastern
relationships between system power reliability and system Turkey. Energy Conversion and Management 47, 1281–1297.
configurations on the one hand and system costs and sys- Esen, H., Inalli, M., Esen, M., 2007. A techno-economic comparison of
tem configurations on the other hand, have been studied. ground-coupled and air-coupled heat Pump system for space cooling.
The optimal configurations of the hybrid system are deter- Building and Environment 42, 1955–1965.
Habib, M.A., Said, S., El-Hadidy, M.A., Al-Zaharna, I., 1999. Optimi-
mined in terms of different desired system reliability
zation procedure of a hybrid photovoltaic wind energy system. Energy
requirements (DPSP) and system costs. It is obvious from 24, 919–929.
the obtained results that the TNPC, TAC and BEDA for Iqbal, M., 1983. An introduction to solar radiation. Academic Press,
1 day battery storage are lower than 2 and 3 days for the Canada, ISBN 0-12-373752–4.
2420 A. Kaabeche et al. / Solar Energy 85 (2011) 2407–2420
Kaabeche, A., Belhamel, M., Ibtiouen, R., Moussa, S., Benhadadi, M.R., Notton, G., Lazarov, V., Stoyanov, L., 2010. Optimal sizing of a grid-
2006. Optimisation d’un système hybride (Eolien–Photovoltaı̈que) connected PV system for various PV module technologies. Renewable
totalement autonome. Revue des Energies Renouvelables 9, 199–209. Energy 35, 541–554.
Kaabeche, A., Belhamel, M., Ibtiouen, R., 2010. Sizing optimization of Posadillo, R., López Luque, R., 2009. Evaluation of the performance of
grid-independent hybrid photovoltaic/wind power generation system. three diffuse hourly irradiation models on tilted surfaces according to
Energy 36, 1214–1222. the utilizability concept. Energy Conversion and Management 50,
Karaki, SH., Chedid, R.B., Ramadan, R., 1999. Probabilistic performance 2324–2330.
assessment of autonomous solar-wind energy conversion systems. Prasad, A.R., Natarajan, E., 2006. Optimization of integrated photovol-
IEEE Transactions on Energy Conversion 14, 766–772. taic–wind power generation systems with battery storage. Energy 31,
Kaviani, A.K., Riahy, G.H., Kouhsari, S.H.M., 2009. Optimal design of a 1943–1954.
reliable hydrogen-based stand-alone wind/PV generating system, Tina, G., Gagliano, S., Raiti, S., 2006. Hybrid solar/wind power system
considering component outages. Renewable Energy 34, 2380–2390. probabilistic modeling for long-term performance assessment. Solar
Klucher, T.M., 1979. Evaluation of models to predict insolation on tilted Energy 80, 578–588.
surfaces. Solar Energy 23, 111–114. Wen, J., Zheng, Y., Donghan, F., 2009. A review on reliability assessment
Kolhe, M., Agbossou, K., Hamelin, J., Bose, T.K., 2003. Analytical model for wind power. Renewable and Sustainable Energy Reviews 13, 2485–
for predicting the performance of photovoltaic array coupled with a 2494.
wind turbine in a stand-alone renewable energy system based on Yang, H.X., Burnett, L., Lu, J., 2003. Weather data and probability
hydrogen. Renewable Energy 28, 727–742. analysis of hybrid photovoltaic–wind power generation systems in
Markvar, T., 2000. Solar Electricity, second ed. J. Wiley & Sons. Hong Kong. Renewable Energy 28, 1813–1824.
Markvart, T., 1996. Sizing of hybrid PV-wind energy systems. Solar Yang, H., Lu, L., Zhou, W., 2007. A novel optimization sizing model for
Energy 59, 277–281. hybrid solar-wind power generation system. Solar Energy 81, 76–84.
Mehleri, E.D., Zervas, P.L., Sarimveis, H., Palyvos, J.A., Markatos, N.C., Yang, H.X., Zhou, W., Lu, L., Fang, Z.H., 2008. Optimal sizing method
2010. Determination of the optimal tilt angle and orientation for solar for stand-alone hybrid solar–wind system with LPSP technology by
photovoltaic arrays. Renewable Energy 35, 2468–2475. using genetic algorithm. Solar Energy 82, 354–367.
Nelson, D.B., Nehrir, M.H., Wang, C., 2006. Unit sizing and cost analysis Yokoyama, R., Ito, K., Yuasa, Y., 1994. Multi-objective optimal unit
of stand-alone hybrid wind/PV/fuel cell power generation system. sizing of hybrid power generation systems utilizing PV and wind
Renewable Energy 31, 1641–1656. energy. Journal of Solar Energy Engineering 116, 167–173.