Renewable Energy 90 (2016) 175e187

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Modeling of integrated renewable energy system for electrication of

a remote area in India
S. Rajanna*, R.P. Saini
Alternate Hydro Energy Centre, Indian Institute of Technology, Roorkee, Uttarakhand 247667, India

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 3 September 2015
Received in revised form
24 December 2015
Accepted 29 December 2015
Available online 7 January 2016

Over the years, renewable energy based power generation has proven to be a cost-effective solution in
stand-alone applications in the regions where grid extension is difcult. The present study focused on
the development of models for optimal sizing of integrated renewable energy (IRE) system to satisfy the
energy needs in different load sectors of four different zones considered in Chamarajanagar district of
Karnataka state in India. The objective of the study is to minimize the total cost of generation and cost of
energy using genetic algorithm (GA) based approach. Considering optimization power factor (OPF) and
expected energy not supplied (EENS), optimum system feasibility has been investigated. Based on the
study, it has been found that IRES is able to provide a feasible solution between 1.0 and 0.8 OPF values.
However, power decit occurs at OPF values less than 0.8 and the proposed model becomes infeasible
under such conditions. Customer interruption cost (CIC) and decit energy (DE) for all zones were also
computed to quantify the reliability of the systems.
2015 Elsevier Ltd. All rights reserved.

Keywords:
Genetic algorithm
Off grid electrication
Integrated renewable energy system
Customer interruption cost
Levelized cost of energy

1. Introduction
In a developing country like India, majority of the population
resides in remote rural areas and number of small isolated communities in such areas live without access to electricity from the
power grid [1]. Extension of power grid based supply to these
remote regions is not cost-effective or feasible. Moreover, conventional solutions proved by diesel generators are costly due to high
cost of fuel and maintenance and have drawbacks such as loud
noise and more importantly, emissions of green-house gases.
Renewable energy sources are clean and may prove to be good
alternative solutions for supplying the required electrical load. Due
to the uncertainty in the availability of renewable energy resources,
it is preferable to use an integrated renewable energy (IRE) system
to increase system reliability [2]. An IRE system utilizes two or more
locally available renewable energy based systems and may be a cost
effective solution to meet the energy demand of remote areas.
Further, such IRE systems are non-polluting, reliable and reduce the
total operating and maintenance cost [3]. In order to utilize the
available renewable energy resources efciently and economically,
optimal models are required to be developed to meet the demand

* Corresponding author. Tel:+918791340700

E-mail address: srajannamce@gmail.com (S. Rajanna).
http://dx.doi.org/10.1016/j.renene.2015.12.067
0960-1481/ 2015 Elsevier Ltd. All rights reserved.

through these resources. However, the modeling of an IRE system is

a very complex task which requires the development of mathematical models for each components [4].
The optimization of integrated renewable energy systems has
been addressed in number of studies. Ramakumar et al. [5] developed a linear programming (LP) approach to design IRES model by
utilizing various renewable energy resources. They minimized the
total annual cost under the constraints of availability of power and
energy requirements. Joshi et al. [6] proposed a decentralized energy planning for three villages in Nepal. The optimization aimed at
minimizing cost function considering mix of energy resources and
conversion devices. Akella et al. [7] formulated an optimal size of
IRE system for Jaunpur block of Uttarakhand state of India. They
used electric power delivery factor was used for optimization
problem to calculate cost of energy from different energy resources
contribution for the considered area. Ashok [8] developed a numerical iterative algorithm to optimal sizing of a PV-wind-microhydro hybrid system for the electrication of rural areas in Western Ghats (Kerala), India. Iniyan et al. [9] developed an optimal
renewable energy model (OREM) for renewable energy allocation
in India for the year 2020e21. They minimized the ratio of cost and
efciency. Ferrer et al. [10] used integer or binary variables to dene
for location and size of the equipment. Khatod et al. [11] developed
systematic analytical approach to minimize the production cost of
small autonomous power system. The study carried out on

176

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Nomenclature
ASPV
BGG
BMG
COE
CIC
CN
CRF
Cumecs
D
DE
DIE
EBGG
EENS
EMHP
EBMG
EWTG
ESPV
EO
FC
g
OTCG
hnet
HT
IEAR
INR
IR

area of SPV (m2)

biogas generator
biomass gasier system
cost of energy (Rs/kWh)
customer interruption cost (Rs in thousands)
curve number
capital recovery factor
cubic meter per second
duration in which load is not meet out (hour)
total energy demand (kWh)
decit in energy (kWh)
annual energy output of biogas generator (kWh)
expected energy not supplied (kWh)
annual energy output of micro hydro generator (kWh)
annual energy output of biomass generator (kWh)
annual energy output of wind turbine generator (kWh)
annual energy output of solar photovoltaic system
(kWh)
total energy production (kWh)
fuel cost (Rs)
acceleration due to gravity (m/s2)
optimal total cost of generation (Rs in thousands)
net head (m)
solar radiation availability (W/m2)
interruption energy assessment rate (Rs/kWh)
Indian rupee
interest rate (%)

uncertainties associated with solar irradiance, wind speed, demand

and outages of various generating units for the period of one year.
Ramakumar et al. [12] developed a design tool IRES-KB based on
the knowledge-based system and reported scenarios based models
for typical remote rural village in India. They minimized the total
capital cost for the proposed IRE system at a pre-selected reliability
level. They also found optimal sizes of energy storage systems to
fulll the energy requirements at the desired reliability level.
Koutroulis et al. [13] investigated optimal number and type of
system components for 20-years using genetic algorithm. They
minimized the total system cost subject to the constraint of the load
power with zero load rejection. Also they compared the total system cost with conventional optimization methods.
Hakimi andTafreshi [14] developed a particle swarm optimization (PSO) based intelligent method applied to the problem of
sizing in a hybrid power systems for residential area. They optimized the net present cost of all the systems. Fatemeh and Rihy [15]
formulated optimized cost function to minimize the net present
cost of hybrid generation system operation over 20 years under the
economic and technical constraints. They computed reliability index from components failure, including wind turbine, PV array,
battery and inverter failure using PSO algorithm. Deshmukh and
Ramakumar [16] proposed two probabilistic models for wind
electric conversion system. The rst one was based on wind speed
distribution while the other was on Markov approach model. They
discussed the inuence of the major parameters involved on the
overall system reliability when operated in parallel with conventional utility grid. Maleki and Fathollah [17] used reliability concept
to maximize the allowable loss of power supply probability (LPSP
max) for the hybrid system. Also, they proposed cost effective
hybrid system and system rank for different LPSP values (0%, 0.3%

kBGG
kBMG
kMHG
kSPV
kWTG
L
Lw
LCE
MHG
N
O&M
OPF
Pi
PSPV
PMHP
PBGG
PBMG
PWTG
Q
SPV
TCG
WTG
hMHG
hBGG
hBMG
hSPV

rw

annual capacity factor for BGG

annual capacity factor for BMG
annual capacity factor for MHP
annual capacity factor for SPV
annual capacity factor for WTG
average annual power load (kW)
length of the watershed (m)
levelized cost of energy (Rs/kWh)
micro hydro generator
life of the plant (yr)
operation and maintenance cost (Rs/kWh)
optimization power factor
capital cost (Rs)
power output of solar photovoltaic system (kW)
power output of micro hydro generator (kW)
power output of biogas generator (kW)
power output of biomass generator (kW)
power output of wind turbine generator (kW)
discharge (m3/s)
solar photovoltaic system
total cost of generation
wind turbine generator
efciency of micro hydro generator (%)
biogas conversion efciency (%)
biomass conversion efciency (%)
Conversion efciency of SPV (%)
density of water (kg/m3)
density of water (kg/m3)

and 1%) and found the most cost effective hybrid system for the
LPSP value of 5%. Amer et al. [18] proposed a hybrid renewable
energy system to satisfy the load demand through searching algorithm for reducing the levelized cost of energy with a satisfactory
range of the production. Alireza and Leandro [2] developed three
grid independent hybrid renewable energy systems for electrication to a small load area in Kerman Iron. They found PV/WT/
battery hybrid system is most cost effective and reliable for meeting
the energy demand of the proposed area.
Venkat and Mccalley [38] presented scenario based nationalwide transmission overlay planning over 40-years through NETPLAN co-optimization tools. They evaluated net benets in terms of
total investment and production cost system. Further, various end
effect mitigation models were also used to simulate a 40-year longterm infrastructure planning problem at yearly time steps [39].
Trishna Das et al. [40] demonstrated the generic bulk energy storage dispatch model for production costing simulation and assess
the economic viability of storage system under different system
scenarios.
Further, levelized cost of energy and levelized unit electricity
cost were considered as an economic sub model for the economic
assessment of IRE systems [19e22]. Similarly loss of power supply
probability, expected energy not supply, loss of load deciency and
loss of load expectation were considered as reliability sub models
for the analysis of IRE system reliability [23e26].
Based on the literature review, it is found that, a few studies
have been carried out to develop modeling of off-grid integrated
renewable energy system by combining resources such as micro
hydro, solar, wind, biomass and biogas. The models developed so
far mainly focus on one or two resources such as solar and wind,
and with battery system for meeting load demand of individual

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

rural village. Very few studies are available for electrication of

cluster/block/zone in remote rural areas, considering objective
function, as economic and reliability indicators, and system size
optimization. Further, the micro hydro power modeling has been
developed based on ow rate (direct runoff depth) determined on
the basis of storm rainfall depth using curve number method. The
present paper deals with the size optimization of IRE models
comprised of micro hydro, solar, wind, biomass and biogas
considering the economic, reliability worth and cost of energy
(COE). Further, the ow rate estimation based on storm rainfall
depth using curve number method. Finally, keeping in view the
accuracy level in results, genetic algorithm is used for modeling of
integrated renewable energy system for the proposed study areas.
2. Case study
Based on the comprehensive assessment on un-electried villages in the Karnataka state, Chamarajanagar district has been
found to be one of the most appropriate remote rural areas for the
present study. The objective of the study is to minimize the total
cost of generation and cost of energy of IRE for the proposed study
areas. Zone wise demography data of chamarajanagar district is
given in Table 1.
2.1. Study area
Karnataka is a state in South West India, which was formed on
November 1956 with the passage of the States Reorganization Act.
Originally known as the State of Mysore, it was renamed Karnataka
in 1973. It comprises 30 districts and covers an area of 191,976 km2
or 5.83% of the total geographical area of India. It is the seventh
largest state by area with 61,130,704 inhabitants according to the
2011 census. Chamarajanagar is the southernmost district in the
state of Karnataka located at 12
000 0000 north latitude and
77
1706000 E east longitudes. The total area of the district is 5101 km2
with 4 blocks: Kollegal, Gundlupet, Chamarajanagar and Yalandur.

177

Chamarajanagar majorly consists of hilly terrain and the plain area

is more fertile than the plateau. Most of the area of the district is
under forest cover which plays an important role not only ecologically, but also economically.
There are 513 small, scattered, and thinly populated villages in
the district. Villagers have small landholdings; terrace cultivation is
the only option and the economy relies largely on agriculture and
animal husbandry. All the villages are mainly dependent on natural
resources. Biomass, as a resource, is available in considerable
amounts and used for cooking as well as heating purposes. The
geographical map of Chamarajanagar is shown in Fig. 1. The area
consists of 428 revenue villages and 85 non-revenue villages. Out of
85 non-revenue villages, 26 are un-electried and considered for
electrication. The villages are categorized into four zones based on
availability of energy resources and population density. The total
number of households in all four zones is 2240 while the population across all zones is 4440. Furthermore, based on the survey it
has found that distance between the proposed zones are more than
40 km and building transmission (micro grid) between these zones
may be economically feasible because of dispersed population,
rugged terrain, and dense forest with wild animals.

3. Load assessment
The total energy demand of all the sectors in the area has been
estimated on the basis of the number of households and population
of the villages. Seasonal load demand was determined for different
load sectors and details of load for all the four zones are given in
Table 2. The electricity demand was mainly constituted by different
sectors such as domestic load (lighting, TV, fan and radio/music
system), community (primary health center, street lights and
school lighting), commercial (lighting for shops and our mill) and
small scale industrial load (saw mill or paddy huller). The estimation of total energy requirement of the villages was made on the
basis of minimum desirable seasonal load prole and hourly load.
The annual energy generation has been calculated using

Table 1
General information of study area.
Features

Zone 1

Zone 2

Zone 3

Zone 4

Name of gram
panchayath

Gopinatham

P.G. Palya

M.M. Hills

Punajanur

Populations (Nos.)
Literates (Nos.)
Illiterates (Nos.)
Marginal
agricultural
labors (Nos.)
Marginal workers
(Nos.)
Marginal
cultivators
(Nos.)
Electried villages
(Nos.)
Un-electried
villages (Nos.)
Households (Nos.)
Household
populations
(Nos.)
Latitude and
Longitude of the
area
Economic status
Social aspects

4092
2028
2064
1021

5588
2554
3034
405

12,645
6342
6303
1039

4200
1945
2255
469

1243

335

2399

679

911

42

495

330

11

23

24

10

302
604

395
750

763
1526

780
1560

12
30 N, 77
410 E

11
530 N,77
170 E

11
590 N,77
80 E

11
460 N, 77
60 E

Zone is economically poor

Lack of education, Communication
Water supply and back ward areas

Very poor
Back ward and weaker zone
back ward and weakest zone
Lack of welfare centre,
Lack of education no proper road and no medical facility, proper road and lack
literacy and back ward area back ward with remote area
of education with remote area

178

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Fig. 1. Location of the study area.

mathematical modeling techniques and annual capacity factor

dened by actual generated energy in kWh/yr to the maximum
generated energy in kWh/yr. The total annual load has been synthesized through HOMER software based on estimated value of
seasonal hourly load. The total energy demand of zone 1, zone 2,
zone 3 and zone 4 was found to be 272398 kWh/yr, 344819 kWh/yr,
516,873 kWh/yr and 673416 kWh/yr respectively. The variation in
seasonal hourly load prole of four zones obtained as shown in
Fig. 2.

4. Resource assessment of study area

In order to assess the potential of renewable energy resources,
an extensive survey was conducted to collect the formation
regarding the availability of biomass, solar irradiation, micro hydro
and wind speed. The potential assessment of solar photovoltaic
(SPV), micro hydro generator (MHG), biogas generator (BGG),
biomass generator (BMG), and wind turbine generator (WTG) was
made as per the available standard methodology [27e32]. Village
wise energy potential in the study area and optimum system sizes

Zone 4

249.6
624
31.2
0
7.2
0.42
0
0.72
0
14
28.08
0.12
0
0.64
160
60
1175.98

Zone 3

244.16
610.4
30.52
0
7.2
0.42
0
0.72
0
16
27.36
0.12
0
0.64
160
60
1157.54
126.4
316
15.8
0
7.2
0.42
0
0.72
0
12
14.04
0.12
0
0.48
120
40
653.18

Zone 2
Zone 1

96.64
241.6
12.08
0
7.2
0.42
0
0.72
0
10
10.8
0.12
0
0.4
100
20
499.98
249.6
624
31.2
327.6
14.4
0.42
0.54
0.72
3.6
14
28.08
0.12
0.18
0.64
160
60
1515.1
244.16
610.4
30.52
320.46
14.4
0.42
0.54
0.72
3.6
16
27.36
0.12
0.18
0.64
160
60
1489.52
96.64
241.6
12.08
126.84
14.4
0.42
0.54
0.72
3.6
10
10.8
0.12
0.18
0.4
100
20
638.34
1560
780
780
780
1
3
3
12
12
7
78
1
1
8
8
3
1526
763
763
763
1
3
3
12
12
8
76
1
1
8
8
3
604
302
302
302
1
3
3
12
12
5
30
1
1
5
5
1
20
100
10
30
600
20
30
20
30
1000
30
20
30
20
5000
5000
CLF lights
TV
Radio
Fan
Refrigerator
CLF lights
Fan
CLF Lights
Fan
Motor
Street light
CLF Lights
Fan
CFL lights
Flour mill
Saw mils

8
8
4
14
24
7
6
3
10
2
12
6
6
4
4
4

8
8
4
0
12
7
0
3
0
2
12
6
0
4
4
4

790
395
395
395
1
3
3
12
12
6
39
1
1
6
6
2

126.4
316
15.8
165.9
14.4
0.42
0.54
0.72
3.6
12
14.04
0.12
0.18
0.48
120
40
830.6

Zone 4
Zone 3
Zone 2
Zone 1

Energy utilization in Summer (kWh)

Zone 4
Zone 3
Zone 2
Zone 1

Numbers in used

Winter

Operating hours
Power rating (watts)
Appliances

Small Industrial Load

Total load in kWh

Shops
Commercial Load

Panchayath hall

Pumping water

High school

Health centre
Community Load

179

are computed and discussed as follows:

4.1. Mathematical model for MHG

Micro hydro turbine generator is able to operate under low head
and low volumetric ow rate conditions, and is suitable for small
water streams. Based on the survey conducted, Zone 1, zone 3 and
zone 4 are found to be sufcient water streams and required ow
rate for the estimation of micro hydro power where as in zone 2 the
hydro potential has found to be no sufcient water streams are exit
in this zone. Therefore, zone 1, zone 3 and zone 4 are considered for
discharge measurement of micro hydro power. The discharge was
estimated for the considered zones on the basis of topography and
monthly average rain fall data obtained from nearby rain gauge
station. Areas, length of watershed and elevation difference of the
stream are computed from the topo-sheet (maps) of the
geographical area. These data were analyzed for desired frequency
and discharge was calculated using CN (curve number) method,
developed by soil conservation service (SCS) of USA in 1969 for
estimating the direct runoff depth based on storm rainfall depth.
Fig. 3 shows the stream ow duration curve for zones 1, 3 and 4.
Actual runoff is calculated by using following Equations [28,29];

Qd
S

I  0:2S2
I 0:8S

when I > 0:2S

(1)

25400
 254
CN

Where Qd is the direct surface runoff depth (mm); I is the monthly

rainfall (mm); S is the maximum potential retention in the watershed (mm); CN is the runoff curve number for hydrological sail
(CN 40e58 for dense forest).
The peak rate of runoff or discharge of a watershed is calculated
by using following Equations;

p
TP 0:6TC TC
r
0:77
.
TC 0:0195 L3w Hnet

Domestic Load

Load sectors

Table 2
Details of load in different sectors of all the 4 zones.

Summer

Energy utilization in winter(kWh)

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

0:0208  A  Qd
TP

(2)

(3)

Where; Q is the rate of runoff or discharge (m3/s); TP is the time to

peak runoff (hour); TC is time of concentration (hour); LW is the
length of the watershed (m); Hnet is the difference in elevation
between the highest point and the outlet point dened as head of
the watershed (m); A is the area of the watershed (hectare).
Using curve number method, the ow rate for zone 1 is estimated as 546 l/s (40% dependability), zone 3 as 821 l/s (50%) and
zone 4 as 775 l/s (40%). The power output of micro hydro power
(MHG) plant has been calculated based on the dependability, ow
rate and head. The generated output of micro hydro power plant as
calculated by using following Equation [30];

PMHP 9:81  Q  hnet  hMHP  rw

(4)

Annual energy generated from a micro-hydro power plant is

calculated by using following Equation;

EMHP PMHP 8760  kMHP

(5)

180

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Fig. 2. Seasonal hourly load prole of four zones.

Fig. 3. Stream ow duration curve.

4.2. Mathematical model for SPV

Based on the longitude and latitude of all the four zones, solar
radiation data has been taken from the HOMER software. Annual
total solar energy potential available for zone 1, zone 2, zone 3 and
zone 4 are estimated as 1944 kWh/m2/yr, 1919 kWh/m2/yr,
1975 kWh/m2/yr and 1925 kWh/m2/yr respectively. The hourly
solar energy output of the SPV generator can be calculated using
following equations [31];

PSPV hSPV  HT  ASPV

PWTG

8
<

0
for
aV 3 bPr ; for
:
Pr
for

0  V  Vcutin and V > Vcutout

Vcutin  V  Vrated
Vrated  V  Vcutout

(6)

(8)

(7)

Where Pr is the rated electrical power; Vcut-in is the cut-in wind

speed; V cut-out is the cut-out wind speed; Constants a and b are the
function of rated speed and cut in speed of the wind turbine, which
are calculated by using following equations;

The annual energy is calculated by following equation;

ESPV PSPV 8760  kSPV

1, zone 2, zone 3 and zone 4 is estimated as 1097 kWh/m2/yr, 2730

(kWh/m2/yr), 2410 (kWh/m2/yr) and 2816 kWh/m2/yr respectively
for electricity generation. The number of aero turbines for all the
zones is computed. The output of wind energy potential in each
zone is calculated as [32];

4.3. Mathematical model for WTG

Based on the wind speed data, wind energy density for different
wind speeds has been computed. The wind energy density for zone

Pr
3
3
Vrated
 Vcutin

(9)

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

3
Vcutin
3
3
Vrated
 Vcutin

5. Model formulation for different resource

(10)

Annual energy of WTG is computed by using following equation

EWTG PWTG 8760  kWTG

(11)

An IRE model has been developed to generate electricity, which

can be used in domestic, community, commercial and small scale
industrial energy sectors of the area as detailed under section- 2.
Optimization model with constraints has been formulated to
minimize the total cost of generation using genetic algorithm as
given below;

4.4. Mathematical model for BGG

Min TCG

Biogas can be produced from livestock manure. In order to

assess the biogas potential, number of cows, buffaloes, goat and
sheep were considered. Based on the survey, the estimated potential of biogas energy in zone 1, zone 2, zone 3 and zone 4 is
estimated as 115 m3/day, 267 m3/day, 264 m3/day and 379 m3/day
respectively for electricity generation. The operational period of
BGG is considered as 10 h per day. The biogas power and energy are
calculated by the following equations [27];

subjected to
P
ERij  DEi

PBGG

 

Total gas yield m3 day  calorific valueBG  hBGS

860  operating hours per day

(12)

The annual energy of biogas generator is calculated by the

following equation;

EBGG PBGS 8760  kBGG

(13)

4.5. Mathematical model for BMG

To estimate biomass potential, forest foliage is considered as
main energy resource for the production of electricity. Collection
efciency of 75% and average collected wood waste of 160 kg/ha/yr
were considered in the biomass estimation. The estimated potential of biomass of zone 1, zone 2, zone 3 and zone 4 considered as
103 tons/yr, 228 tons/yr, 224 tons/yr and 340 tons/yr respectively
which can be made available for electricity generation. In the present study, biomass based gasication through gasier has been
considered. The producer gas produced from the biomass is
considered as biofuels. The duration of system operation is
considered as 12 h per day. The power and energy generated is
calculated by using following expression [31];

PBMG

181

LCEj  ERij

(16)

P ERij
 RAj
hij

(17)

ERij  0
Where; TCG is total generation cost (Rs); LCE is levelized cost of
energy (Rs/kWh) for the jth resources; ERij is optimal amount of
energy of jth resources options for ith end use (kWh); DEi is the
total energy demand for ith end use (kWh); RAj is the availability of
the jth resources options for ith end use (kWh); hij is the conversion
efciency for the jth resources options for ith end use.

5.1. Economic criteria

Modeling of GA considers values of cost parameters and technical parameters related to different sources. Cost parameter
comprises capital cost, maintenance cost, plant life and fuel cost of
the system components [33,34]. The values of cost parameter
considered under the present study are given in Table 3. The
technical parameters considered as input to GA as; considered sizes
based on mathematical modeling, operating hours of biomass and
biogas generators and capacity factors of system components. The
technical parameter of various system components considered for
different zones are given in Table 4.
For economic based modeling, the levelized cost of energy has
been considered as an economic indicator and cost of energy have
been calculated based on the cost parameters of the system components. The levelized cost of energy and capital recover factor of
each system component has been computed by using following
equations [35];

Total fuel wood availabilityton=yr  calorific valueBM  hBMG  1000

365  860  operating hours per day

Annual energy of BMG is calculated by the following equation;

EBMG PBMG 8760  kBMG

(15)

Using mathematical models discussed above, potential of all

considered renewable energy resources has been estimated.
Biomass resource has the maximum potential followed by micro
hydro power, wind energy and solar energy.

(14)

Table 3
Cost parameters considered for different systems under the present study [33,34].
Cost parameters

Units

Systems
MHP

SPV

WTG

BGG

BMG

Capital cost
O & M cost
Plant life
Interest rate
Fuel Cost
LCE

Rs in thousands/kW
% capital cost
Yrs
%
Rs
Rs/kWh

85
3
25
11
0
2.32

125
3
30
11
0
8.0

175
4
25
11
0
5.53

40
4
20
11
2.8
4.55

50
4
15
11
3
4.71

182

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Table 4
Technical parameters of various system components considered for different zones.
Technical Parameters

Zones

Power output of MHP

Annual capacity factors of MHP
SPV modules (240Wp)
Annual capacity factors of SPV
WT Generators (1 kW Rating)
Annual capacity factors of WTG
Power output of BGG
Annual capacity factors of BGG
Power output of BMG
Annual capacity factors of BMG

Units

Zone 1

Zone 2

Zone 3

Zone 4

kW
%
Nos.
%
Nos.
%
kW
%
kW
%

24
58.63
13
25.0
6
54.16
17
41.66
22
50

e
e
25
25.0
6
58.33
35
45.83
51
50

30
66.84
34
25.0
9
58.33
39
41.66
50
50

28.5
66.84
58
25.0
12
54.16
50
45.83
76
50

2
P 3
fCRFi; n O&Mg Pi
6
7
i
P
LCE 4
5 FC
87:6 ki

(18)


CRF


i1 in
1 in  1

generated energy in kWh=yr

system rating in kW  365  24

(19)

(20)

5.2. Reliability criteria

The reliability model has been developed based on the concept
of optimization power factor (OPF) and expected energy not supply
(EENS), which are considered as a probability reliability index. The
developed model was optimized using genetic algorithm. EENS was
considered a probabilistic index to assess reliability of the renewable energy system. EENS represents expected energy not supplied
when load surpasses available generation. The equation used for
calculating EENS is given below;

EENS L:D

Under the present study, annual energy estimation has been

considered in the model formation and additional investments in
peaking generation capacity to meet peak loads are not considered.
6. Genetic algorithm based optimization
Genetic algorithm is a articial intelligence technique for solving both constrained and un constrained optimization problems. At
each iteration, GA performs a series of computations on the current
population to produce a new generation; selection rules, crossover
rules, and mutation rules [37]. GA was used to work out the optimization problem and developed approach is implemented in
MATLAB environment.
For solving the optimization problem, objective function with
constraints, variable and constant have been dened in M-le
(MATLAB code). The energy system components such as MHG, SPV,
WTS, BGG and BMG are considered as ve independent variables.
Genetic algorithm toolbox has been used to determine the tness
function of the optimization problem. It is found that precise results
with a minimum number of generations were considered as
specied functions. The crossover, mutation and selection processes were carried out only on successful member components.
A random generation of population members (possible solutions) by GA led to the generation of the initial population. Every
possible solution is a decision vector code having upper and lower
constraints. The initial generation is developed in the course of
consecutive iterations and the tness function is estimated by
evaluating member components of each generation. Fig. 4 shows
the owchart to explain the genetic algorithm for the optimization
process. Although a population size of 50 was found adequate for a
lower degree of optimization of the system and it has been
observed that 80 number of generations gave the optimal solution.
In a majority of cases however, the requisite number of generations
came out to be smaller than 60. The crossover and the mutation
function have been selected as 0.8 and 0.2 respectively. The selection and crossover functions were the Roulette and the arithmetic
one respectively. The mutation function was selected to be the
constraints dependent and the stopping criteria are the number of
specied generations. Other functions may be used depending on
the optimization problem.

(21)
7. Results and discussions

Where L is the average annual power load (kW); D is duration in

which load is not meet out (hour). The OPF can be stated as an
energy index of reliability, which can be dened as ratio of decit
energy to total energy demand of the system (kWh) and is
expressed as;

OPF

EO  EENS
EO

(22)

The reliability cost should correspond with the system's worth

or benet to society, where reliability worth can be measured
through the interception cost. For calculating reliability worth, the
customer interruption costs (CIC) related to the respective reliability level need to be estimated. CIC can be calculated by using
following equation;

CIC EENS  IEAR

(23)

Where interrupted energy assessment rate (IEAR) is the total

average value of COE of a completely reliable system.
Billinton and Allan [36] directed the association between
optimal total cost and optimum reliability of the system. Customer
interruption cost has been computed for different OPF values by
considering EENS and IEAR.

Optimization results in terms of optimum total generation cost

and cost of energy considered based on OPF value for all the zones
have been obtained through GA based approach. Furthermore,
decit energy and corresponding CIC have been computed for
different OPF values based on probability index in terms of EENS.
The results of the proposed approach are discussed below;
7.1. Cost of energy for 1.0 OPF
Based on the genetic algorithm approach, zone wise unit cost of
energy has been computed for OPF value 1.0 and results are given in
Table 5. For OPF value 1.0, the cost of energy for zone 1, zone 2, zone
3 and zone 4 were found as Rs 4.09/kWh, Rs 4.50/kWh, Rs 3.91/
kWh and Rs 4.26/kWh respectively. It has been found that the
percentage energy contribution of BMG is highest, followed by
MHP and BGG. However, SPV and WTG have less energy contribution in all the zones. System cost was found to be different for all
zones as shown in Fig. 5.
For OPF value of 1.0 (100% energy supplied from all resources
with no decit in energy), the unit cost of energy has been found to
be Rs 4.10/kWh while COE found to be as Rs 3.97/kWh for OPF value
of 0.8 (20% is the decit in energy, only 80% of the load is supplied

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

183

Fig. 4. Flow chart for a GA based approach.

Table 5
Optimization results in terms of optimal cost with COE for 1.0 OPF.
Zones MHP
Contribution SPV
Contribution WTG
Contribution BGG
Contribution BMG
Contribution OTCG(Rs in Thousands) COE (Rs./kWh)
(kWh/yr) (%)
(kWh/yr) (%)
(kWh/yr) (%)
(kWh/yr) (%)
(kWh/yr) (%)
Zone
Zone
Zone
Zone

1 97,071
2 e
3 210,652
4 147,571

35.64
e
40.76
21.92

10,943
13,139
13,388
29,660

4.02
3.81
2.59
4.14

13,463
12,136
14,016
28,131

4.94
3.05
2.71
4.17

58,402
96,178
92,667
170,638

from all resources). Fig. 6 shows the variation of energy contribution with OPF values from different system components. It has been
observed from Fig. 6 that the IRE model is capable of supplying
energy up to 80% of the load as long as the OPF value is 0.8. For OPF
value less than 0.8, the system becomes infeasible due to the
unavailability/non-functioning of resources. Hence the cost of
generation is increased. Therefore, it is concluded that, decit in
energy increases with an increase in customer interruption cost
and the proposed IRE model is found to be unfeasible for OPF value
below 0.8.
In case of OPF values are in the range of 0.8 and 0.2, energy
contribution from MHG and BGG increases considerably. However,
contribution from SPV and wind decrease even though sufcient

21.44
27.89
17.93
25.34

92,519
223,366
186,150
297,416

33.97
64.78
36.01
44.16

1110
1550
2020
2870

4.10
4.5
3.91
4.26

potential is availability (Table 6). In such cases, the resources are

inadequate to meet energy demands. Further, high cost of energy of
SPV system restricts its use. Therefore, based on the energy distribution of resources in the area, either a suitably designed and
optimized IRE system, or a biomass based gasier engine system
may be used for electrication of the area under study.
Fig. 7 shows the variation in optimal cost of energy with OPF
value for different zones. Feasible and unfeasible regions for all
zones have been predicted based on the cost of decit in energy by
varying OPF values from 1.0 to 0.2. This implies that feasibility solutions in terms of fullling energy requirements of each zone are
found for OPF value range of 1.0 to 0.8. For OPF values 1 means 100%
of energy served by resources and no decit in energy (ie., unmet

184

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Fig. 5. Optimal cost of the system components for 1.0 OPF.

Fig. 6. Variation of energy contribution with OPF values from different system components.

Table 6
Energy contribution with OPF values from different system components.
Zones

OPF

MHG

SPV

WTG

BGG

BMG

Zone 1

1
0.8
0.6
0.4
0.2
1
0.8
0.6
0.4
0.2
1
0.8
0.6
0.4
0.2
1
0.8
0.6
0.4
0.2

97,071
76,127
33,477
18,105
20,514
e
e
e
e
e
210,652
156,210
139,526
98,822
80,518
147,571
119,537
95,751
42,407
40,243

10,943
8536
1348
0
0
13,139
11,925
6086
4150
0
13,388
10,766
7456
0
0
29,660
25,821
10,252
3033
0

13,463
10,420
2221
2291
1273
12,136
12,898
10,651
7488
876
14,016
12,285
9158
2886
0
28,131
26,222
11,265
6512
0

58,402
53,524
34,426
29,997
5022
96,178
72,967
84,569
62,164
9517
92,667
82,077
59,178
13,246
18,471
170,638
138,531
108,286
35,138
32,655

92,519
69,312
91,967
58,566
27,671
223,366
178,065
105,585
64,126
58,571
186,150
152,160
94,806
91,794
35,695
297,416
228,621
178,496
182,276
61,785

Zone 2

Zone 3

Zone 4

energy is zero), while OPF value 0.8 means 80% of energy served by
resources and 20% of energy is not served (unmet energy is 20%)
due to the unavailability/non-functioning of resources. For OPF

value below 0.8, the decit in energy increases signicantly with

decrease in energy fullling requirements results in the system
unfeasible. Therefore, as OPF decreases, Optimized cost is also decreases with increasing decit in energy that means, unmet energy
increases with decreasing optimized cost of energy.
In order to compute decit in energy and corresponding CIC,
different OPF values have been determined by considering the
probability index in terms of EENS. Obtained results are shown in
Figs. 8e9. It has been found that decit with the decrease in OPF
values result in increase in customer interruption cost. In other
words, energy demands are not fullled due to inadequate energy
generation.
Based on the study, biomass gasier based power generation is
considered as a major energy sources for electricity generation in
all the zones. The total emission of CO2 and corresponding carbon
price for all the system components has been calculated based on
the CO2 emission(ton) for 1kWh energy [41e43]. Based on the
estimated values, the CO2 emission from BMG is more as compared
to other resources such as SPV, WTG, BGG and MHG. The estimated
value of CO2 emission in zone 1, zone 2, zone 3 and zone 4 are
92.42 tons/yr. 223.14 ton/yr, 185.96 tons/yr and 297.11 tons/yr and
the corresponding carbon price of respective zones are 9.24 Rs/ton,
22.31 Rs/ton, 18.59 Rs/ton and 29.71 Rs/ton respectively. The zone
wise variation in carbon price vs. emission is shown in Fig. 10. The
obtained results are based on the annual energy estimation. Apart

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

185

Fig. 7. Variation in optimal cost of energy with OPF value for different zones.

Fig. 8. Variation in decit in energy with OPF value for different zones.

Fig. 9. Variation in customer interruption cost with OPF value for different zones.

from present analysis, the system might need additional investments in peaking generation capacity to meet peak loads, since
they have not explicitly modeled for peak hours under the present
study.

8. Conclusions
In this paper, an optimal IRE model comprising MHG, SPV, WTG,
BGG and BMG has been developed for four different zones in the
Chamarajanagar district of Karnataka, India. Zone wise optimized
total cost and COE were determined through GA based approach.

186

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

Fig. 10. Variation in carbon price vs.CO2 emission.

Based on the results, COE obtained through GA has found to be Rs.

4.10/kWh, Rs. 4.50/kWh, Rs. 3.91/kWh and Rs. 4.26/kWh for zone 1,
zone 2, zone 3 and zone 4 respectively. Zone wise decit energy and
customer interruption cost were worked out for different values of
OPF through reliability based concept in terms of EENS and OPF. It
has been found that the proposed model can provide feasible solution if the OPF value lies within the range of 1.0 to 0.8. The results
of this paper indicate IRES can be feasible for this region. The results
of this paper indicates IRES can be feasible for considered region.
However, further investigations are required to evaluate other
competing options such as transmission expansion between zones,
coal retrots, energy storage and demand response, in order to
perform a comprehensive planning analysis and estimate cost and
emission impacts.
Acknowledgments
The rst author express sincere thanks to Alternate Hydro Energy Centre and Quality Improvement program (QIP) Centre, IIT,
Roorkee, India for providing research facilities and All India council
for Technical Education (AICTE), Government of India for providing
nancial assistance in the form of research fellowship. He also express heartedly gratitude to Malnad College of Engineering (Govt
aided), Hassan, Karnataka, India for nancially sponsoring him to
pursue doctoral study.
References
[1] Wai Kean Yap, Vishy Karri, An off-grid hybrid PV/diesel model as a planning
and design tool, incorporating dynamic and ANN modeling techniques,
Renew. Energy 78 (1) (2015) 42e50.
[2] Alireza Askarzadesh, S.C. Leandro dos, A novel framework for optimization of
a grid independent hybrid renewable energy system: a case study of Iran, Sol.
Energy 112 (1) (2015) 383e396.
[3] R. Ramakumar, Energizing rural areas of developing countries using IRES,
Proceding IECEC 3 (1996) 1536e1541.
[4] Antonio Cano, J. Francisco, S. Higinio, M. Fernandez, Optimal sizing of standalone hybrid systems based on PV/WT/FC by using several methodologies,
J. Energy Inst. 87 (1) (2014) 330e340.
[5] R. Ramakumar, S.P. Sudhakara, K. Ashenayi, A linear programming approach
to the design of integrated renewable energy systems for developing countries, IEEE Trans. Energy Convers. 1 (4) (1986) 18e24.
[6] Bhatti S. Joshi, N.K. Bansal, Decentralized energy planning model for a typical
village in India, Energy 17 (9) (1992) 869e876.
[7] A.K. Akella, M.P. Sharma, R.P. Raini, Optimum utilization of renewable energy
sources in a remote area, Renew. Sustain Energy Rev. 11 (1) (2007) 894e908.
[8] S. Ashok, Optimized model for community-based hybrid energy system,
Renew. Energy 32 (1) (2007) 1155e1164.
[9] S. Iniyan, L. Suganthi, T.R. Jagadeesan, Renewable energy planning for India in

21st century, Renew. Energy 14 (1) (1998) 453e457.

[10] L. Ferrer-Mart, B. Domenech, A. Garcia-Villoria, R. Pastor, A MILP model to
design hybrid windephotovoltaic isolated rural electrication projects in
developing countries, Eur. J. Oper. Res. 226 (1) (2013) 293e300.
[11] D.K. Khatod, Vinay Pant, J.D. Sharma, Analytical approach for well-being
assessment of small autonomous power systems with solar and wind energy sources, IEEE Trans. Energy Convers. 25 (2) (2010) 535e545.
[12] R. Ramakumar, M. Abouzahr, K. Ashenay, A knowledge-based approach to the
design of integrated renewable energy systems, IEEE Trans. Energy Convers.
7 (4) (1992) 648e659.
[13] E. Koutroulis, K. Dionysian, P. Antonis, K. Kostas, Methodology for optimal
sizing of stand-alone photovoltaic/wind-generator systems using genetic algorithms, Sol. Energy 80 (2006) 1072e1088.
[14] S.M. Hakimi, S.M. Moghaddas-Tafreshi, Optimal sizing of a stand-alone hybrid
power system via particle swarm optimization for Kanauji area in South-East
of Iran, Renew. Energy 34 (2009) 1855e1862.
[15] J.A. Fatemeh, A. Riahy, Design of an optimum hybrid renewable energy system
considering reliability indices, Proc. of ICEE Conf. (2010).
[16] R.G. Deshmukh, R. Ramakumar, Reliability analysis of combined wind electric
and conventional generation systems, Sol. Energy 28 (4) (1982) 345e352.
[17] M. Akbar, P. Fathollah, Optimal sizing of autonomous hybrids photovoltaic/
wind/battery power system with LPSP technology by using evolutionary algorithms, Sol. Energy 115 (2015) 471e483.
[18] Namaane A. Motaz, N.K. Sirdi, Optimization of hybrid renewable energy systems HRES using PSO for cost reduction, Energy Procedia 42 (2013) 318e327.
[19] R. Ramakumar, N.G. Butler, A.P. Rodriguez, S. Mani, Economic aspects of
advanced energy technologies, Proc. of ICEE Conf. (1993).
[20] S. Diaf, D. Diaf, M. Belhamel, M. Haddadi, A. Louche, A methodology for
optimal sizing of autonomous hybrid PV/Wind system, Energy Policy 35
(2007) 5708e5718.
[21] Hongxing Yang, Lin Lu, Wei Zhou, A Novel Optimization sizing model for
hybrid solar-wind power generation system, Sol. Energy 81 (1) (2007) 76e84.
[22] B.S. Bogdan, Z.M. Salameh, Methodology for optimally sizing the combination
of a battery bank and PV array in a wind/PV hybrid system, IEEE Trans. Energy
Convers. 11 (2) (1996) 367e375.
[23] B.S. Bogdan, Z.M. Salameh, Optimum photovoltaic array size for a hybrid
wind/PV system, IEEE Trans. Energy Convers. 9 (3) (1994) 482e488.
[24] A.B. Patil, R.P. Saini, M.P. Sharma, Development of IREOM model based on
seasonally varying load prole for hilly remote areas of Uttarakhand State in
India, Energy 36 (2011) 5690e5702.
[25] A.M. Al-ashwal, I.S. Moghram, Proportion assessment of combined PV-wind
generating systems, Renew. Energy 10 (1) (1997) 43e51.
[26] G. Tina Wang, Chanan Singh, Multicriteria design of hybrid power generation
systems based on a modied particle swarm optimization algorithm, IEEE
Trans. Energy Convers. 24 (1) (2009) 163e172.
[27] A. Chauhan, R.P. Saini, A review on integrated renewable energy system based
power generation for stand-alone applications: Congurations, storage options, sizing methodologies and control, Renew. Sustain Energy Rev. 38 (1)
(2014) 99e120.
[28] A.M. Michael, T.P. ojha, Principles of Agricultural Engineering, Jain Brothers,
2013.
[29] K. Subramanian, Engineering Hydrology, McGraw-Hill Companies, 2012.
[30] A. Gupta, R.P. Saini, M.P. Sharma, Modeling of hybrid energy system Part I:
Problem formulation and model development, Renew. Energy 36 (1) (2011)
459e465.
[31] A. Gupta, R.P. Saini, M.P. Sharma, Steady-state modeling of hybrid energy
system for off grid electrication of cluster of villages, Renew. Energy 35 (1)
(2010) 520e535.

S. Rajanna, R.P. Saini / Renewable Energy 90 (2016) 175e187

[32] A. Kaabeche, M. Belhamel, A sizing optimization of grid-independent hybrid
photovoltaic/wind power generation system, Energy 36 (1) (2011)
1214e1222.
[33] MNRE, development of renewable energy, Role of proper incentive mechanism.
Available<http://mnre.gov.in/le-manager/UserFiles/presentationspwc-workshop-06092012/IREDA.pdf > [accesses 19.07.15].
[34] Saini RP. Chauhan, Renewable energy based off grid rural electrication in
Uttarakhand State of India: technology option, modeling methods, barriers
and recommendations, Renew. Sustain Energy Rev. 51 (1) (2015) 662e681.
[35] R. Ramakumar, W. Huges, Renewable energy sources and rural development
in developing countries, IEEE Trans. Educ. E-24 (3) (1981) 242e251.
[36] R. Billiton, R.N. Allan, Reliability Evaluation of Power System, Pitman Books,
January 1983.
[37] M.S. Ismail, M. Moghavemia, M.I. Mahlia, Genetic algorithm based optimization on modeling and design of hybrid renewable energy systems, Energy
Convers. Manage 85 (1) (2014) 120e130.

187

[38] V. Krishnan, J.D. Mccalley, Santiago Lemos, James Bushnell, National-wide

transmission overlay design and benets assessment for the U.S, Energy Pol.
56 (2013) 221e232.
[39] V. Krishnan, J.D. Mccalley, Building foresight in long-term infrastructure
planning using end-effect mitigation models, IEEE Syst. J. (2015).
[40] T. Das, V. Krishnan, J.D. Mccalley, Assessing the benets and economics of bulk
energy storage technologies in the power grid, Appl. Energy 139 (2015)
104e118.
[41] Sovacool D. Nugent, K. Benjamin, Assessing the lifecycle greenhouse gas
emissions from solar PV and wind energy: a critical meta-survey, Energy
Policy 65 (2014) 229e244.
[42] Bond Tighe, Expanded environmental notication form, Russell Biomass Proj.
(2005).
[43] Wolfgang Bauer, Stefan Bauer, Thomas Bauer, Energy and greenhouse gas
analysis for biogas power plants, Int. Conf. Renew. Energies Power Qual.
(2012).