You are on page 1of 15

Renewable and Sustainable Energy Reviews 50 (2015) 755–769

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews


journal homepage: www.elsevier.com/locate/rser

Review of recent trends in optimization techniques for solar


photovoltaic–wind based hybrid energy systems
Sunanda Sinha, S.S. Chandel n
Centre for Energy & Environmental Engineering, National Institute of Technology, Hamirpur 177005, Himachal Pradesh, India

art ic l e i nf o a b s t r a c t

Article history: An update literature review on trends in optimization techniques used for the design and development
Received 14 January 2015 of solar photovoltaic–wind based hybrid energy systems is presented. The main objective is to identify
Received in revised form latest promising techniques for the optimization of solar photovoltaic (PV)–wind based hybrid systems.
11 April 2015
Different techniques used by researchers for the optimization of renewable based hybrid energy systems
Accepted 12 May 2015
Available online 31 May 2015
are reviewed along with PV–wind based hybrid system sizing methodology, is presented. Optimization
studies during last 2.5 decades by researchers using traditional and new generation methods are
Keywords: analyzed and sixteen optimization methods including hybrid algorithms are presented. The trend shows
Hybrid energy systems that new generation artificial intelligence algorithms are mostly used during last decade as these require
Solar–wind based hybrid systems
less computation time and have better accuracy, good convergence in comparison to traditional
Solar photovoltaics
methods. The study suggests using hybridization of two or more algorithms to overcome the limitations
Wind turbine
Optimization techniques of a single algorithm. Additionally some other techniques are identified for follow up research in the
Hybrid algorithms design of PV–wind hybrid systems. This review will be useful for researchers to face complexity and
challenges in renewable energy based hybrid system research.
& 2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 756
2. PV–wind hybrid system sizing methodologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1. Requirements for PV–wind hybrid system optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1.1. Meteorological data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1.2. Load profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1.3. System configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1.4. Energy system model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
2.1.5. Optimization results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
2.2. Criteria for PV–wind hybrid system optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
2.2.1. Reliability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
2.2.2. Cost analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
2.3. Modeling of hybrid system components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760
2.3.1. Modeling of photovoltaic system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
2.3.2. Modeling of wind generator system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
2.3.3. Modeling of battery system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
2.4. Conditions and steps for the execution of a hybrid optimization problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
3. Optimization techniques used in PV–wind based hybrid research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
3.1. Traditional approach for optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
3.1.1. Graphical construction technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
3.1.2. Iterative techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
3.1.3. Probabilistic approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762

n
Corresponding author. Tel.: þ 91 1972 254748; fax: þ91 1972 223834.
E-mail addresses: sschandel2013@gmail.com, chandel_shyam@yahoo.com (S.S. Chandel).

http://dx.doi.org/10.1016/j.rser.2015.05.040
1364-0321/& 2015 Elsevier Ltd. All rights reserved.
756 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

3.1.4. Trade-off approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762


3.1.5. Linear programming technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
3.2. New generation approach for optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763
3.2.1. Genetic Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763
3.2.2. Particle swarm optimization (PSO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
3.2.3. Simulated annealing (SA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
3.2.4. Other new generation approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
3.2.5. Hybrid algorithm optimization techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766
5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 767
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 767

Nomenclature CRF ¼ capital recovery factor


Tc ¼ cell temperature
V ci cut-in wind speed ηc ¼ charge efficiency (depends on the SOC and the char-
V co cut out wind speed ging current and has a value between 0.65 and 0.85)
Ib direct normal solar radiations Cs ¼ cost of solar PV system
P f aliure load which cannot be served within a time period ηdis ¼ discharge efficiency (generally taken equal to 1)
Pw power output from wind turbine generator. σ ðtÞ ¼ hourly self-discharge rate depending on the battery
ηPV PV system efficiency state (taken constant at about 0.02%)
Pr rated power of a wind turbine i¼ interest rate
Vr rated speed of the wind turbine C bat ¼ nominal capacity of the battery (A h).
α power law exponent Pw ¼ power generated from wind turbine
P total total load T proj ¼ project life time
CB cost of battery bank APV ¼ PV system area
C other cost of other systems and accessories Tr ¼ reference temperature for cell efficiency
Cw cost of wind system C ann ¼ total annualized cost
Id diffuse solar radiations P load ¼ total load to be delivered or load demand
ηw efficiency of wind turbine generator and correspond- P total ¼ total power generated from resources
ing converter, COE cost of energy
Edef icit energy deficit within a certain time period (t) DOD depth of discharge,
P load load demand during a period LA level of autonomy
ηm module efficiency LCC life cycle cost
ηpc power conditioning equipment efficiency LLP/LOLP loss of load probability
T f ailure power failure time period LOLH loss of load hours
Ps power generated from solar PV system LOLR loss of load risk
IT solar radiation on a tilted surface LPSP loss of power supply probability
Rd tilt factor for the diffused solar radiation. NPC net present cost
Rr tilt factor for the reflected solar radiation. p cumulative probability of meteorological status which
H total total hours of operation corresponds to electrical energy generation
H LOL total number of hours during which loss of load q probability of failure
(LOL) occurs SOC state of charge
Aw total swept area SPL system performance level
C total total system cost UL unmet load
T total total working time V and V0 wind speeds at heights h and h0 (h0 is the reference
β¼ array efficiency temperature coefficient height).

1. Introduction resource availability at a particular location, but in the present


study we have focused only on PV–wind based hybrid systems as
The fossil fuel resources are becoming scarce due to ever solar and wind are most promising power generating sources due
increasing energy demand in commercial, industrial, agricultural to their complementary nature advantage. Wind speeds are often
and domestic sector. In this context, alternative energy resources low in periods when the solar resource is at its best. On the other
like solar, wind, biomass, bio-fuel, hydro and geothermal etc. are hand, the wind is often stronger in seasons when there are less
being utilized largely to generate power in recent years. A renew- solar resource. But these sources depend on climatic conditions
able energy based hybrid system offers a better option than a which are unpredictable thus making the design of a hybrid
single source based system in terms of cost, reliability and system complex. In order to improve the reliability of a PV wind
efficiency. One or more energy sources can be utilized in renew- hybrid system other sources like fuel cell, diesel generator can also
able energy based hybrid systems (REHS) which can work as stand be integrated. Such a hybrid system can meet the basic power
alone or in a grid connected mode. Different types of hybrid requirements in a non-electrified remote area where grid power is
system combinations are feasible depending on the need and not available. Besides this our main research focus in the present
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 757

study is on PV wind based hybrid system applications in urban, hybrid systems as discussed in a comprehensive review by Sinha
rural and remote regions. However, the techniques covered are and Chandel [1]. However, these softwares have some limitations
also useful for studying other types of hybrid systems. like black box coding, different working platforms, unavailability
A solar–wind hybrid system consists of PV array, wind turbine, of some of the softwares and are also not as flexible as optimiza-
battery bank, inverter, controller, and other accessories. The tion techniques which can be used as per research criteria.
schematic is shown (Fig. 1).The generated power from solar and Sizing and optimization techniques must efficiently search for
wind energy charge the battery after meeting load demand but if an optimum combination of parameters like system cost, system
generated power is less than demand then battery will supply the reliability, PV system size, tilt angle of PV panels, battery size, wind
load as per storage capacity. turbine size with hub height because, over sizing causes higher
The steps to be followed for the efficient design and planning of system costs and under sizing causes insufficient power supply. A
a PV–wind based hybrid system, are shown in Fig. 2. number of authors have carried out detailed reviews on sizing and
There are some constraints in the formulation and solution of optimization techniques [2–10]. Summary highlights of these
the designing and optimization approach like resource availability, studies are given in Table 1.
technology, efficiency, mathematical models and many more. But The paper is organized as follows: Section 2 provides an
advancement in computational techniques has made it easy to overview of PV–wind hybrid system sizing methodologies;
deal with optimization problems by using a number of optimiza- Section 3 gives a literature survey on optimization techniques
tion and simulation techniques. A number of simulation tools used in PV–wind based hybrid system research; Section 4 deals
HOMER (Hybrid Optimization Model of Electric Renewable), with discussion on the recent trends of optimization followed by
HYBRID2 (The Hybrid Power System Simulation Model), HOGA Conclusion in Section 5.
(Hybrid Optimization using Genetic Algorithm) etc. are used for
optimizing, designing and performance evaluation of PV–wind
2. PV–wind hybrid system sizing methodologies

In this section sizing methodologies, requirements, criteria,


mathematical modeling, conditions and execution process for a
PV–wind based hybrid system are discussed.

2.1. Requirements for PV–wind hybrid system optimization

Fig. 1. Schematic of a solar–wind hybrid system. The input parameters required for a PV–wind hybrid system
optimization are as follows.

2.1.1. Meteorological data


Analysis of meteorological characteristics of the location has to be
made for optimization process. Measured solar and wind resource
data are the main inputs for PV wind hybrid based system optimiza-
tion. The time series measured minute wise, hourly or daily weather
data are preferable. In case measured data are not available for the
location then satellite based data or estimated data can also be used
for preliminary studies if accuracy is not the major consideration.

2.1.2. Load profile


Yearly electric load demand profile is one of the necessary steps
to design-planning and optimization of a hybrid system. It is
difficult to find out and also complicated to analyze real load
demand with all minute fluctuations, therefore hourly or daily
averages of load demand is generally used for design-optimization
purpose. But to have a real or nearest to real load variations for all
the seasons is really difficult task, if it fails then system designed
can be oversized or undersized.

2.1.3. System configuration


After the prefeasibility studies based on weather data (e.g. wind
speed, solar irradiation and temperature) and load demand the
selection of proper sizing of equipment can be made. But this sizing
process must be according to nature of PV and wind, e.g. if the study
location have good solar potential than wind then the hybrid system
must be configured with maximum share of PV system and minimum
share of wind system.

2.1.4. Energy system model


Energy system models are the mathematical models developed to
represent various energy-related problems reliably. These models are
used to identify and solve problems using various computing
Fig. 2. Basic steps for renewable energy based hybrid system design and planning. systems. The accuracy of the developed models for PV/wind system
758 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

Table 1
Summary highlights of review studies on sizing and optimization of renewable energy based hybrid systems.

Reference/ Systems studied Topics covered Highlights


year of
study

Zhou et al. Only standalone PV–wind hybrid systems  Criteria for optimizations and simulation  Artificial intelligence techniques are identified to
[2] modeling of photovoltaic system, wind energy be promising which requires to be further
system, battery storage system explored
 Software tools for hybrid solar–wind system
reviewed are
○ HOMER
○ HYBRID2
○ HOGA

 Sizing methods reviewed are


○ Graphic construction method
○ Probabilistic approach
○ Iterative technique
○ Artificial Intelligence method
○ Multi Objective optimization

Erdinc and Covers all types of renewable energy based  Software tools discussed are: HOMER, HYBRID2, Promising techniques identified are:
Uzunoglu hybrid systems. The General Algebraic Modeling System (GAMS),  Ant colony algorithm
[3] Optimization of Renewable Intermittent Energies  Artificial immune system algorithm
with Hydrogen for Autonomous Electrification  Tabu Search
(ORIENTE), OptQuest, LINDO, WDILOG2, Dividing  Honey Bee Mating Algorithm
Rectangles (DIRECT), Determining Optimum  Bacterial Foraging Algorithm
Integration of RES (DOIRES),Simulation of  Game Theory
Photovoltaic Energy Systems (SimPhoSys), Geo-
Spatial Planner for Energy Investment Strategies, Hybrid methodologies require to be improved
Grid-connected Renewable Hybrid Systems further
Optimization (GRHYSO), H2RES
 Optimization techniques discussed are:
○ Genetic algorithm
○ Particle swarm optimization
○ Simulated annealing
○ Linear programming
○ Simplex algorithm
○ Neural Networks
○ Evolutionary algorithm
○ Stochastic, iterative, probabilistic,
parametric and numerical approaches

Fadaee and Study covers PV–battery, PV–wind–battery  Optimization Techniques reviewed:  GA and PSO as most useful and promising multi
Radzi[4] and PV–wind–diesel–battery hybrid systems ○ Genetic Algorithm objective optimization methods in hybrid
○ Honey Bee mating Optimization system design
○ Particle Swarm Optimization
○ Evolutionary Algorithm
○ Artificial Intelligence
○ Pareto-based multi-objective optimization
and parallel processing

Luna-Rubio Study includes all types of renewable energy  Reviewed hybrid system performance indicators  Study suggests hybrid optimization
et al. [5] based hybrid systems (Loss of power supply probability (LPSP), methodologies are superior to other methods.
Levelized cost of energy (LCE))
○ Hybrid energy system sizing
methodologies reviewed are:
○ Probabilistic methods,
○ Iterative methods,
○ Hybrid methods (Genetic Algorithm,
Artificial Intelligence)
○ Analytical methods including commercial
software tools and/or numerical
approximations of system component)

Khatib et al. Study includes standalone and grid  Standalone PV systems size optimization  Artificial intelligence techniques have the
[6] connected PV systems, PV–diesel generator methods reviewed are: potential to improve the process of optimization.
systems, PV–wind systems, PV–wind–diesel ○ Intuitive methods
generator systems ○ Numerical methods
○ Analytical methods
○ Other methods (Artificial Intelligence)
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 759

Table 1 (continued )

Reference/ Systems studied Topics covered Highlights


year of
study

 Reviewed grid connected hybrid systems sizing


optimization methods:
○ Intuitive methods
○ Numerical methods
○ Artificial intelligence methods

Upadhyay Covers all types of renewable energy based  Discussed design parameters, evaluation criteria  Hybrid optimization methodologies are
and hybrid systems and control and energy management of hybrid recommended for hybrid system research to
Sharma energy systems avoid the limitation of one methodology
[7]  Software tools reviewed are:
○ HOMER
○ HOGA
○ RETScreen
○ HYBRIDS
○ TRNSYS

 Sizing methodologies reviewed are:


○ Graphic construction methods
○ Probabilistic methods
○ Analytical methods
○ Iterative methods
○ Artificial intelligence methods
○ Hybrid methods

Chauhan Covers all types of renewable energy based  Reviewed various types of integrated  Concludes artificial intelligence techniques as a
and Saini hybrid systems configurations (DC coupled, AC coupled, Hybrid better technique than other deterministic
[8] DC–AC coupled) and various options for energy methods
storage technologies with system control and  Genetic algorithm (GA), harmony search (HS),
management particle swarm optimization (PSO), biogeography
 Mathematical model for wind, Micro-hydro, based optimization (BBO) are the most promising
solar, biomass gasifier energy systems are studied Algorithms in future research
 Reviewed various sizing methodologies used in
hybrid system study
○ Artificial intelligence
○ Multi objective design
○ Iterative approach
○ Analytical method
○ Probabilistic approach
○ Graphical construction method

 Software tools reviewed are:


○ HOMER
○ HOGA
○ RET Screen
○ HYBRIDS
○ TRNSYS

Bourennani Focused on solar–wind–fuel cell based Multi-Objective Optimizations (MOO) for Hybrid Article concludes that hybrid energy system
et al. [9] hybrid systems Energy System Design requires more interaction between both energy and
MOO research
Bhandari Study covers PV–wind–diesel–battery hybrid  Reviewed energy scenario and various types of Study concludes that artificial intelligence may
et al. systems optimal design criteria provide good optimization of system without
[104]  Mathematical model for wind, solar, battery, extensive long term weather data
diesel generator systems are studied
 Reviewed various sizing methodologies used in
hybrid system study
○ Graphical construction method
○ Probabilistic approach
○ Iterative approach
○ Artificial intelligence

 Software based approach: HOMER


760 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

components is important in optimizing. Therefore these models must 2.2.1. Reliability analysis
include all necessary variables which affect energy conversion. These The dependency on nature and unpredictability of solar and
mathematical models should use simple concepts but in many cases wind resources have a great impact on energy production which
the complexity of a model increases or researcher may not include leads to unreliable power supply during cloudy or non windy days.
some important factors which may lead to inaccuracy. A system is reliable if it is able to supply required power to the
electrical load within a specific time period. A power reliability
analysis is essential for system design and optimization process.
2.1.5. Optimization results There are several methods in literature to calculate the reliability
Optimization results must be accurate enough to avoid excess of a hybrid energy system which are summarized in Table 2.
or deficit power and it is only possible if above four steps are
followed. Although power generation from PV–wind system is
2.2.2. Cost analysis
location dependant process, yet generalization of optimization
Cost analysis of a hybrid system is important for optimization so as
results for nearby location is also important.
to deliver energy at minimum cost. There are several methods used to
calculate cost of a hybrid energy system like net present cost, life cycle
2.2. Criteria for PV–wind hybrid system optimization cost and cost of energy which are summarized in Table 3.

A PV–wind hybrid system has to be optimally designed to deliver 2.3. Modeling of hybrid system components
power for a particular load demand reliably. The optimum hybrid
system configuration must satisfy and compromise between two main The performance of a hybrid solar–wind system is dependent
objectives: power reliability and system cost. on its components. The mathematical modeling equations of a PV

Table 2
Methods to determine reliability of hybrid energy systems.

Sl. Reliability criteria/method Definition and representation Reference


no.

1 Loss of power supply Most widely used method in which probability of insufficient power supply to load demand is taken into account while [11]
probability (LPSP) designing the hybrid system. LPSP is the ratio of power supply deficits to the electric load demand during a certain period
Pn
Edef icit
LPSP ¼ t P
¼ 1
n
P load
t ¼ 1

2 Loss of load probability Defined as the power failure time period divided by the total working time of the hybrid system [11]
(LLP/LOLP)
3 Unmet load (UL) The load which cannot be served divided by the total load of a time period (normally one year) [11]
Pn
P f ailure
UL ¼ t P
¼ 1
n
P total
t ¼ 1

4 System performance level SPL is defined as the probability of unsatisfied load [12]
(SPL)
5 Loss of load hours (LOLH) LOLH is the summation of loss of load expectation in hours over a specified time (usually one year) that the power system [13]
is unable to meet load requirements due to lack of power at an instant excluding the effects of component breakdown or
maintenance time
6 Loss of load risk (LOLR) LOLR is defined as the probability of the generating system failure to meet the daily electrical energy demand due to [14]
deficient energy of the renewable energy sources used
LOLR can be can be represent as
LOLR ¼ 1  p Or LOLR ¼ q
7 Level of autonomy (LA) LA deals with two main parameters namely the total number of hours in which loss of load (LOL) occurs and the total [5]
hours of operation. If autonomy increases system will be more reliable but same time cost will be higher
LA ¼ 1  HHtotal
LOL

Table 3
Cost analysis methods of hybrid energy system.

Sl. Cost analysis Definition and representation Reference


no. method

1 Net present cost NPC reflects cost of energy for a particular system and is defined as the ratio of the total annualized cost of the system to the annual [2]
electricity delivered by the system. NPC is total present value of cash flows including initial cost of system components, replacement
cost of any component within project lifetime, and cost of maintenance, i.e. investment costs plus the discounted present values of
all future costs during system’s lifetime
C ann
NPC ¼ CRFði;T proj Þ

2 Life cycle cost LCC deals with sum of all recurring and one-time (non-recurring) costs over the full life span or a specified period of a good, service, [2]
structure, or system
Includes purchase price, installation cost, operating costs, maintenance and upgrade costs, and remaining (residual or salvage)
value at the end of ownership or its useful life
3 Cost of energy COE reflects the cost of energy or electricity generation and as the ratio of total annualized cost of the system to the annual [2]
electricity delivered by the system. Total annualized cost includes all the costs over the system’s lifetime from initial investment and
capital costs, to operations and maintenance (e.g. fuel) and financing costs
COE¼ total cost/energy produced
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 761

system, wind system and battery system are described in this During charging process
section.
SOCðt þ 1Þ ¼ SOCðtÞ:½1  σ ðtÞ þ½I bat ðtÞ:Δt:ηc ðtÞ=C bat  ð9Þ
During discharging process
2.3.1. Modeling of photovoltaic system
The power output from PV system with area Apv (m2) is given SOCðt þ 1Þ ¼ SOCðtÞ:½1  σ ðtÞ ½I bat ðtÞ:Δt=ηdis ðtÞ=C bat ð10Þ
by:
With ð1  DODÞ rSOCðtÞ r1
P ¼ I T APV ηPV ð1Þ
where IT is expressed as 2.4. Conditions and steps for the execution of a hybrid optimization
problem
I T ¼ I b Rb þ I d Rd þ ðI b þI d ÞR ð2Þ
The PV system efficiency is expressed as The two main conditions to be fulfilled for a good designed
hybrid system are maximum power reliability and minimum cost.
ηPV ¼ ηm ηpc ½1  βðT c  T r Þ ð3Þ
Total energy generated by wind and PV systems can be
expressed as
2.3.2. Modeling of wind generator system P total ¼ P w þ P s ð11Þ
Power output of wind turbines for a location depends on wind
The total energy generated may or may not satisfy the load
speed at hub height which can be calculated using power-law
demand and three different situations may arise depending upon
equation given below
the load demand and total power generated which are shown as:
 α
V h
¼ ð4Þ i. P total ¼ P load
V0 h0
Total energy generated matches the load demand, so there is
Total power available from wind turbine is given by: no excess and deficiency in power.
P ¼ P w Aw ηw ð5Þ ii. P total o P load
Total energy generated is less than the load demand, so there
Different wind turbines have different power output and perfor-
will be a deficiency in power and battery will have to supply
mance curves. Therefore, the modeling equation of a wind system is
the load.
strongly influenced by the power curve of the wind turbine used.
iii. P total 4 P load
Fig. 3 shows a typical wind turbine power curve characteristics and
using this curve power output Pw (kW/m2) from wind generator can
Total energy generated is greater than the load demand, so
be calculated as follows
8 excess power is generated and battery will be charged.
>
< 0; if V o V ci and V 4V co The optimum combination of a solar–wind hybrid system can
Pw ¼ aV 3  bP r ; if V ci oV o V r ð6Þ make the best compromise with system cost. The economical
>
:P ;
r if V r oV o V co approach is composed of minimization of the costs of individual
components like PV, battery, wind system etc. This cost includes
where a and b are the co-efficients are given as capital, maintenance and replacement cost of various components
Pr of the system.
a¼  ð7Þ The total cost can be expressed as
V 3r  V ci
C total ¼ fC s þ C w þ C B þ C other gmin ð12Þ
V 3ci The optimization process is shown in flow chart (Fig. 4).The
b¼   ð8Þ
V r  V 3ci
3 first step of the optimal sizing methodology consists of providing
meteorological data–electric load demand and system input data.
The second step consists of mathematical equations with relia-
2.3.3. Modeling of battery system bility and cost conditions. The next step is using optimization
Battery is used to store surplus generated energy, to regulate methods fulfilling the criteria of system configuration.
system voltage and to supply load in case of insufficient power
generation occurs from the hybrid system. Battery sizing depends
3. Optimization techniques used in PV–wind based hybrid
on maximum depth of discharge (DOD), temperature and battery
research
life. A battery’s state of charge (SOC) is expressed as follows:

The power supplied by solar and wind resources mainly depends


on meteorological conditions so it is necessary to maximize the power
during sunny or windy days with properly optimally sized design.
There is a continuously increasing interest in the development of
hybrid energy systems using various optimal sizing techniques. The
classification of optimization algorithm can be carried out in a number
of ways but in this paper the algorithms are divided into two simplest
categories namely traditional approaches and new generation
approaches. A traditional approach follows a rigorous procedure for
example linear programming, iterative techniques etc. whereas new
generation approaches are genetic algorithm, particle swarm
optimization etc.
In this section an overview of various optimization techniques
Fig. 3. Wind turbine power curve. used in PV–wind hybrid systems are described.
762 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

(LEC) and life cycle unit cost (LUC) of power generation with battery
bank. Borowy and Salameh [22] determined optimum size of a battery
bank and PV array for a stand-alone hybrid wind–PV system and
concluded that the optimum mix depends on the particular site, load
profile, and the desired reliability of the hybrid system. Ashok [23]
proposed an iterative method where an optimal hybrid system was
obtained among different renewable energy combinations including
PV and wind technology ensuring system’s reliability and minimum
life cycle cost which is applicable to renewable power generation in
any rural village.
Iterative optimization method requires more computational
efforts and usually two main parameters PV module tilt angle
and wind turbine tower height are not optimized in most of the
reported studies.

3.1.3. Probabilistic approach


A probabilistic model is that where there are multiple possible
outcomes, with varying degrees of certainty or uncertainty of
occurrence. Probabilistic approach considers the effect of random
variability of parameters. Probabilistic approaches enable variation
and uncertainty to be quantified, mainly by using distributions
instead of fixed values. This technique is also used by some
researchers to solve hybrid system sizing problem. However, the
Fig. 4. General execution process followed in optimal sizing of hybrid systems.
probabilistic approach cannot represent the dynamic changing
performance of the hybrid system which is main disadvantage of
3.1. Traditional approach for optimization this method.
Bagul et al. [24] used probabilistic approach to determine the
A number of studies have been carried out using like graphical relation between number of photovoltaic arrays and batteries to
construction method, iterative techniques etc which are meet a given reliability if wind turbine rated capacity and load
discussed below. demand are known. Karaki et al. [25] used a probabilistic treat-
ment of an autonomous solar–wind energy conversion system
3.1.1. Graphical construction technique delivering a load. The methodology used fix an upper limit on
Borowy and Salameh [15] presented a graphical construction required storage batteries and predicts the expected energy not
technique for finding optimum combination of a stand-alone PV– supplied (EENS) to the hybrid system considering charging/dis-
wind hybrid system based on long-term meteorological data. For a charging cycles of the batteries. Combination of the two separate
typical load consumption of a house in Massachusetts and a desired PV and wind models were obtained and convolution theory was
LPSP, the optimum configuration of battery bank and PV array was used. Tina et al. [26] used probabilistic approach for the long-term
calculated based on the minimum cost of the system. Markvart [16] performance assessment of a solar–wind hybrid system for stand-
used graphical technique to optimally design a solar–wind hybrid alone and grid-linked applications and developed an analytical
power generation system using monthly-average solar and wind model. The authors modified the model developed by Karaki et al.
energy values. This is a basic and easily understandable method with [25] using energy index of reliability (EIR). The validity of the
no complexity, but this technique is not used currently as it is not developed probabilistic model is shown and results are compared
flexible and is based on various approximations. with those obtained from time-series simulation.

3.1.2. Iterative techniques 3.1.4. Trade-off approach


An iterative method is a mathematical procedure that generates This approach is not widely used in hybrid system sizing and
approximate solutions for problems. This is a recursive process which not much literature is available on this method. Chedid et al. [27]
stops when the best configuration is reached as per specifications. In used trade-off/risk method and presented a decision support
hybrid renewable energy research, iterative approach is used from technique to study the design of a hybrid solar–wind power
beginning to design and optimize. Yang et al. [17] proposed a Hybrid system for grid-connected applications based on simultaneous
Solar–Wind System Optimization (HSWSO) model, using iterative maximization of reliability and minimization of cost. Gavanidou
optimization technique. For power reliability and system cost, authors and Bakirtzis [28] applied the method in the design of a standa-
considered LPSP and Levelized Cost of Energy model with parameters lone system. The outcome of the method is not a unique “optimal”
like capacity of PV system, rated power of wind system, and capacity design, but a small set of robust designs and the final decision is
of the battery bank. An iterative optimization method was used by left to the decision makers. The main disadvantage of this
Kellogg et al. [18,19] to optimize the wind turbine size and number of approach is that emission control, which has a major influence
PV modules. Diaf et al. [20] used a methodology to perform optimal on the final trade-off curve, is not taken into account.
sizing of an autonomous hybrid PV–wind system among a set of
systems components, which meets the desired system reliability 3.1.5. Linear programming technique
requirements and lowest levelized cost of energy. The linear programming method was first developed by Leonid
Prasad and Natarajan [21] used an iterative technique to determine Kantorovich in 1939 and is a widely used technique for sizing and
the optimum size of solar panels, wind turbines and capacity of optimization of renewable systems. Chedid and Rehman [29]
batteries of a hybrid system based on deficiency of power supply proposed an optimal design of wind–solar hybrid system using this
probability (DPSP), relative excess power generated (REPG), unutilized technique to minimize the cost of electricity while meeting the load
energy probability (UEP), life cycle cost (LEC), levelized energy cost requirements in a reliable manner along with considering
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 763

environmental factors. Huneke et al. [30] used linear programming global optimum of the problem, cannot assure constant optimiza-
to obtain optimal configuration for a solar–wind–battery–diesel tion response times etc. A number of researches have used the
based power generator combination for two real off-grid energy application of GA for the optimal design and operation of PV–wind
systems in India and Colombia. The optimization results for both based hybrid energy systems.
studies show the feasible combination of PV–battery and diesel An optimal sizing of standalone PV–wind systems is proposed
generator. Nogueira et al. [31] used the methodology for sizing and by Koutroulis et al. [37] using GA to select the optimal number of
simulation of a PV–wind–battery hybrid energy system and linear units with minimum cost, subject to load demand fulfillment. In
programming with minimum cost and high reliability. Lee et al. [32] another study Koutroulis et al. [38] presented a GA based optimal
formulated a Linear programming based new optimization model sizing of desalination systems by PV–wind generators as a power-
for hybrid power systems considering various power losses and supplied unit. Yang et al. [39–40] used GA to optimize the
studied three case studies. The main focus was not on minimizing configurations of hybrid solar–wind–battery bank system where
the total cost of the system but on minimizing the outsourced the decision variables are the number of PV modules, wind
electricity supply and electricity storage capacity. Saif et al. [33] turbines and batteries, the PV slope angle and wind turbine tower
formulated a problem of a PV–wind–diesel–battery hybrid power height. This method was proposed for a hybrid system which
system as a Linear Programming model with two objectives: supplies power for a telecommunication relay station. Bilal et al.
minimizing total cost and minimizing total CO2 emissions, while [41] proposed an optimized sizing of a hybrid solar–wind–battery
capping the Expected Unserved Energy (EUE). Nagabhushana et al. system through multi-objective genetic algorithm satisfying two
[34] used LP to calculate the sizes of the components of a PV–wind principle aims of annualized cost minimization and minimization
hybrid renewable energy system for three locations in Karnataka. of the loss of power supply probability (LPSP). Nafeh [42] used GA
Among all the techniques discussed above, linear programming to yield optimum PV wind, and battery ratings with minimum cost
technique is found to be best than other approaches as it improves the and power reliability. Abdullrahman and Addoweesh [43] pro-
quality of decision. Also linear programming technique is more flexible posed a methodology of optimum sizing of a PV–wind–diesel–
than other methods and a wide range of problems can be solved easily. battery hybrid system using GA including the effect of wind
Comparison of various traditional hybrid system optimization turbine parameters such as rated speed and rated power. The
techniques are summarized in Table 4. decision variables included in the optimization process are the PV
array capacity, wind turbine number, battery bank number and
diesel generator rated power. Atia and Yamada [44] adopted a two
3.2. New generation approach for optimization
layer (main and secondary) genetic algorithm coupled with a local
optimizer known as a Hybrid Genetic Algorithm (HGA) and used in
This approach is widely used now-a-days in renewable energy
designing and controlling of PV–wind–diesel systems. Authors
research to solve optimization and design problems. These are not
also found that the HGA is more powerful algorithm than the
restricted to local optimum configuration but also to determine
conventional GA. Merei et al. [45] used GA to optimize PV–wind–
global optimum system configuration with relative computational
diesel–battery hybrid system with three different battery technol-
simplicity as compared to traditional optimization methods.
ogies. A controlled elitist genetic algorithm has been applied by
Abbes et al. [46] to perform a multi-objective design of PV–wind–
3.2.1. Genetic Algorithm battery hybrid system in order to find the best compromise
Genetic Algorithm (GA) is a search process that mimics the between three objectives: life cycle cost (LCC), system embodied
process of natural selection and was developed by John Holland in energy (EE) and loss of power supply probability (LPSP). Shi et al.
1960–1970 period [35,36].GA generates solutions to optimization [47] used multi-objective genetic algorithm to study techno-
problems using techniques inspired by natural evolution such as economical performance of the PV–wind hybrid energy system
inheritance, mutation, selection, and crossover. GA has several and optimized three objectives e.g. total system cost, autonomy
advantages: it can solve problems with multiple solutions, easy to level, and wasted energy rate with the PV array peak power, the
understand and can easily be transferred to existing simulations wind generator rated power and the rated capacitor of the battery
and models etc. It has some limitations like a tendency to converge as decisive variables. Mostofi and Shayeghi [48] used GA to solve
towards local optima or even arbitrary points rather than the the optimization problem of a PV–wind–hydro–fuel cell hybrid

Table 4
Comparison of various traditional approaches used for PV–wind based hybrid system optimization.

Technique Highlights Strength Weakness Reference

Graphical Presents graphical solution of Easy to understand and use Some important factors (like tilt angle of PV, [15–16]
construction optimization problem wind turbine hub height) are completely
technique neglected in this technique
Iterative approach It is a recursive process which stops Easy to understand; Tracks defects at early Each phase of iteration is rigid with no overlaps [17–23]
when the best configuration is stages
reached as per design specifications
Probabilistic Based on the effects of random Easy to understand and use It cannot represent the dynamic changing [24–26]
approach variability of upon the performance performance of the hybrid system
of an system
Trade-off method Based on a situation that involves Easy to understand Not much used in renewable energy [27–28]
losing one quality or aspect of applications
something in return for gaining
another quality or aspect
Linear programming Based on a mathematical model Best suitable for solving complex problems; Linearity in relation of variables; assumptions [29–34]
represented by linear relationships simple to use; more flexible than any other of linear programming are also unrealistic:
methods; a wide range of problems can be there is a change in relation between input,
solved easily output gain, loss etc.
764 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

system and compared results with HOMER software and con- more favorable than the other PSO variants and PV–battery based
cluded that GA has better accuracy than HOMER. Shadmand and hybrid systems are suitable for most areas of the country due to
Balog [49] presented Multi-Objective Genetic Algorithm (MOGA) the good solar radiation availability and low windy nature.
to determine the PV–wind hybrid system design, optimized by
considering multiple criteria including size, cost, and availability.
3.2.3. Simulated annealing (SA)
Tégani et al. [50] used GA for optimal sizing of PV–wind hybrid
Simulated annealing, which mimics material annealing proces-
system with a lifespan of 20 years. Multi-Objectives Genetic
sing, was developed by Kirkpatrick, Gelatt and Vecchi in 1983 [63].
Algorithm approach is also used by Bilal et al. [51] to design and
It is a trajectory based random search technique for global
optimize a stand-alone hybrid PV/wind/diesel/battery system
optimization. The main advantage of simulated annealing is its
minimizing the LCE and the CO2 emission for Senegal.
ability to avoid being trapped in local minima. Simulated anneal-
ing is a robust and versatile technique which can deal with highly
nonlinear models, chaotic and noisy data and many constraints.
3.2.2. Particle swarm optimization (PSO)
The main weakness of SA is that the quality of the outcome may be
Particle swarm optimization was developed by Kennedy and
poor. Till now little literature has been reported using SA in this
Eberhart [52,53] based on the research of bird and fish movement
field. Ekren and Ekren [64] used simulated annealing (SA) algo-
behavior. The advantages of PSO are: speed of the researching is
rithm for optimizing size of a PV–wind–battery hybrid energy
very fast; calculation in PSO is simple as compared to other
system to minimize total cost. The decision variables are PV size,
methods and can be completed easily. Limitations of this optimi-
wind turbine rotor swept area and the battery capacity used by the
zation algorithm are that it cannot work out the problems of non-
authors and found that SA algorithm gives better result than the
coordinate system, easily suffers from the partial optimism etc.
Response Surface Methodology (RSM).
The use of the method in the PV–wind hybrid system is just
beginning and few literatures are reported till now.
Basir and Sadeh [54] have taken a combination of wind, 3.2.4. Other new generation approaches
photovoltaic and tidal energy with battery source and used PSO Several artificial intelligence based/metaheuristic/heuristic/
to determine the capacity of hybrid system. The Equivalent Loss nature inspired clever algorithms are formulated and used in
Factor (ELF) has been used as an index to evaluate the system recent years. Some of the techniques which are used in PV–wind
reliability level. The paper also compared hybrid wind, solar and hybrid system studies are discussed in this sub section.
battery combination with wind, solar, tidal and battery combina- Ant colony algorithms [65–66] were initially proposed by
tion and found the second combination as more economical. Lee Marco Dorigo in 1992 in his PhD thesis. The algorithm was aiming
and Chen [55] used an evolutionary PSO algorithm to solve wind– to search for an optimal path in a graph, based on the behavior of
photovoltaic capacity coordination with the aim of maximizing the ants seeking a path between their colony and a source of food.
benefit–cost ratio. Kaviani et al. [56] optimized a hybrid wind– A small renewable hydroelectric, wind and solar resources
photovoltaic–fuel cell generation system over its 20 years of based on hybrid hydrogen storage in the northwestern region of
operation with PSO. The aim was to minimize the annual cost of Iran (Ardebil Province) is studied by Menshsari et al. [67]. In this
the hybrid system subject to reliable supply to meet load demand. paper ant colony algorithm is used by authors for improving the
Bansal et al. [57] used Meta Particle Swarm Optimization (MPSO) technical and economic performance of the hybrid system. Xu
to solve the PV–wind–battery hybrid system optimization pro- et al. [68] proposed a specific graph-based ant system to minimize
blem. Using this improved PSO technique local minimum trap can the total capital cost, subject to the constraint of the LPSP for sizing
be avoided thus proving it as an effective technique. Sharafi and of standalone hybrid wind/PV power systems.
ELMekkawy [58] studied PSO simulation based approach to tackle Bacterial Foraging Algorithm (BFO) [69] is inspired by the group
the multi-objective optimization problem for a hybrid system foraging behavior of bacteria such as Escherichia coli and Myx-
consisting of wind turbine, photovoltaic panels, diesel generator, ococcus xanthus. It is the chemo taxis behavior of bacteria that will
batteries, fuel cell, electrolyzer and hydrogen tank. Pirhaghshe- perceive chemical gradients in the environment and move toward
nasvali and Asaei [59] proposed a standalone PV–wind–diesel– or away from specific signals. Bazyar [70] used Bacterial Foraging
battery based hybrid system for Kerman, Iran and used PSO to Algorithm (BFA) for an optimal design of integrated wind–PV–
obtain the optimal sizes for wind turbine, PV system, battery diesel–battery system for supply of power demand in remote and
banks and the diesel generator. Borhanazad et al. [60] used Multi- rural areas of Ardebil, Iran. Result shows that hybrid wind–PV–
Objective Particle Swarm Optimization (MOPSO) in order to obtain diesel–battery is suitable from economical point of view.
the best configuration of the PV–wind–diesel–battery based Artificial bee colony algorithm (ABC) is an optimization algo-
hybrid system for three stations in Iran namely Nahavand, Raf- rithm based on the intelligent foraging behavior of honey bee
sanjan, and Khash. Maleki and Askarzadeh [61] used four heuristic swarm, proposed by Karaboga and Basturk [71,72]. In ABC, the
algorithms namely, particle swarm optimization (PSO), tabu search position of a food source represents a possible solution to the
(TS), simulated annealing (SA), and harmony search (HS) for optimization problem and the nectar amount of a food source
optimum sizing of a cost-effective PV–wind–fuel cell and PV– corresponds to the quality (fitness) of the associated solution.
wind–battery based hybrid systems. The results show that PSO is Nasiraghdam and Jadid [73] presented a novel multi-objective
more robust and promising than other three algorithms used by artificial bee colony algorithm to solve the distribution system
authors. A study to determine the optimum dimensions of hybrid reconfiguration and hybrid (photovoltaic–wind turbine–fuel cell)
photovoltaic systems, wind power, and storage battery bank has energy system sizing. This article also found total power loss, the
been carried out by Maleki et al. [62] in the remote regions of the total electrical energy cost, and the total emission produced by hybrid
South, North-West and North-East of Iran. Authors studied the energy system and grid minimization and the voltage stability index
performance of five different PSO variants and three more algo- (VSI) of distribution system maximization. To optimally size a hybrid
rithms namely tabu search, simulated annealing and harmony energy system based on PV–wind–fuel cell for Rafsanjan, Iran, an
search (HS). PSO, modified PSO (MPSO), PSO based on repulsion efficient artificial bee swarm optimization (ABSO) algorithm is pro-
factor (PSO-RF), PSO with constriction factor (PSO-CF), and PSO posed by Maleki and Askarzadeh [74]. Simulation results shows that
with adaptive inertia weight (PSO-W) these are the variants used PV–wind–fuel cell is the most cost-effective hybrid system and wind–
in the study. It has been concluded by the authors that PSO-CF is fuel cell and PV–fuel cell systems are in the other ranks. Tudu et al.
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 765

[75] studied the optimal sizing combinations of solar–wind–hydro– colonies and imperialists that all together form some empires.
fuel cell based hybrid systems for catering the load of a remote village Imperialistic competition among these empires forms the basis of
of Kerala, India using bee algorithm. Results show that the combina- ICA. During this competition, weak empires collapse and powerful
tion of hydro–wind–fuel cell is the most feasible hybrid energy system ones take possession of their colonies.
in relation to the net present cost for the study location. Maleki and Gharavi et al. [88] determines the optimal sizes of autonomous
Pourfayaz [76] evaluated the performance of different evolutionary and non-autonomous PV–wind–electrolyzer–fuel cell based hybrid
algorithms for optimum sizing of a PV–wind–battery based hybrid system with considerations for economics, reliability indices, and
system which can continuously satisfy a particular load demand with environmental emissions. Authors used fuzzy logic for multi-
minimal total annual cost. Total seven heuristic algorithms, namely, objective problem solving and ICA for optimization purposes. Ranjbar
particle swarm optimization, tabu search, simulated annealing, and Kouhi [89] used GA, PSO and ICA to minimize total cost of PV–
improved particle swarm optimization (IPSO), improved harmony wind–fuel cell based hybrid system and also studied impact of tariff
search (IHS), improved harmony search-based simulated annealing change on the optimal cost of operation for Kerman province, Iran.
(IHSBSA), and artificial bee swarm optimization (ABSO), are applied to The study deals with three different cases where thermal and
the proposed system and the results are compared in terms of total electrical loads are used. Results show that hybrid systems with
annual cost. Authors found that ABSO shows more promising results multiple source units perform better than single source unit and the
than other six algorithm used. entire algorithms have nearly same results.
Bio-geography is the science of studying the behavior of species Tabu Search [90] originally proposed by Glover, which is an
in nature against time and space and species immigration and iterative procedure that starts from a random initial solution and tries
emigration between habitats, which is a probable solution of the to find a better solution escaping local optima. TS used in PV–wind
problem. Biogeography-based optimization (BBO) [77] optimizes a hybrid system studies which is discussed in this paper [61–62,76,91].
problem by maintaining a population of candidate solutions, and
creating new candidate solutions by combining existing ones
according to a simple formula. BBO belongs to the class of 3.2.5. Hybrid algorithm optimization techniques
metaheuristics since it includes many variations, and since it does The hybrid algorithms are developed using advantages and
not make any assumptions about the problem and can therefore disadvantages of two or more optimization methods. The research
be applied to a wide class of problems. Kumar et al. [78] used BBO in this area has dramatically grown up recently.
algorithm to solve the sizing problem of the hybrid PV–wind– Katsigiannis et al. [91] used simulated annealing (SA) and Tabu
diesel–battery system by considering economical and reliability Search (TS), for the solution of autonomous hybrid power system’s
constraints of the system to supply in the area of Jaipur, Rajasthan optimal sizing problem. In the study minimization of cost of energy is
(India). Authors also made a comparison between HOMER soft- the main aim and the design variables are: wind turbines size,
ware, BBO, GA, PSO, comprehensive learning particle swarm photovoltaic system size, diesel generator size, biodiesel generator
optimization (CLPSO) [79] and ensemble of mutation and cross- size, fuel cells size, batteries size, converter size, and dispatch strategy.
over strategies and parameters in DE (EPSDE) algorithm [80] and Authors suggested that in compared to the solutions provided by
shown that BBO algorithm is more rapid and give minimum cost individual SA or TS methods, hybrid SA-TS improved the obtained
as compared to others. Optimal design of a PV–wind–diesel– solutions, in terms of quality and convergence.
battery system is accomplished through BBO for Jaipur, India by Askarzadeh [92] presented an optimum design to determine
Gupta et al [81]. For wind speed and solar radiation forecasting the number of PV panels, wind turbines and batteries so that the
Back propagation trained Artificial Neural Network (BPANN) based total annual cost of the hybrid system subject to some constraints
time-series forecasting methods are implemented by authors and is minimized. Three algorithms are merged namely: chaotic search
concluded that use of forecast data has high influence on optimal (CS), harmony search (HS) and SA to develop a novel discrete
sizing algorithm performance. chaotic harmony search-based simulated annealing algorithm
Geem et al. [82] proposed Harmony Search (HS) algorithm, (DCHSSA). Tutkun [93] investigated a power management for
which is inspired by the improvisation process of jazz musicians. off-grid PV–wind hybrid system using GA and SVM (Support
Optimal sizing of different generation systems (PV–wind–diesel– Vector Machine). In this article binary coded genetic algorithm is
battery, wind–diesel–battery, PV–diesel–battery, and diesel alone) used to minimize the operation cost of this hybrid system and for
are studied by Maleki and Askarzadeh [83] and found that wind– power scheduling the SVM regression method. Khatib et al. [94]
diesel–battery is the most cost-effective system for the study area studied an optimization of hybrid PV–wind system based on loss
Rafsanjani, Iran. Authors used discrete version of harmony search of load probability (LLP) and system cost using hybrid iterative/
(DHS) and discrete simulated annealing (DSA) for the analysis, and genetic algorithm. Authors used the algorithm in two parts firstly a
found that DHS performs better than DSA. set of possible configurations for the proposed system is deter-
Rashedi et al. [84,85] proposed gravitational search algorithm mined by using the iterative part, while in second part the genetic
(GSA) based on Newton’s law of gravitation and the second law of algorithm is applied to find the optimum configuration. Dehghan
motion and superior than the traditional intelligent optimization et al. [95] hybridized PSO and HS algorithm to find optimal sizing
algorithms according to optimization precision and convergence of a hybrid PV–wind plant which can supply the electric load
speed. Wu et al. [86] proposed an enhanced gravitational search demand in a reliable manner and minimum costs. Zhou and Sun
algorithm (EGSA) to optimize the unit output and cost for large- [96] proposed an improved Simulated Annealing Particle Swarm
scale wind–PV–battery storage power generation in Zhangbei, Optimization (SAPSO) algorithm to optimize a wind–solar–bat-
China. Also to demonstrate the superiority of the proposed tery–super capacitor based hybrid system. Basic objective of this
algorithm, a comparative analysis among EGSA, ANN and PSO work is to minimize one-time investment and operation costs
were applied which shows that the unit cost of power generation where the constraints are utilization rate and reliability of power
obtained using EGSA is lower than the unit cost of power supply. Authors also compared the proposed SAPSO algorithm
generation obtained using PSO and ANN. with traditional PSO algorithm and the result shows that new
Imperialist Competitive Algorithm (ICA) is mainly inspired by algorithm is faster than the traditional one and effectiveness of the
imperialistic competition which is developed in 2007 by Atashpaz- new hybrid algorithm is shown. Abdelhak et al. [97] proposed an
Gargari and Lucas [87]. ICA starts with an initial population. optimum sizing methodology for PV–wind–battery hybrid system
Population individuals called countries are divided into two types: by using Fuzzy-Adaptive Genetic Algorithm. This Algorithm is used
766 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

Table 5
Comparison of new generation optimization techniques for PV–wind hybrid systems.

Technique Highlights Strength Weakness Reference

Genetic Mimics process of natural evolution, like Can solve problems with multiple solutions; Convergence speed is slower than other [35–51]
Algorithm inheritance, mutation, selection, and easily transferable to existing simulations and stochastic algorithms; cannot assure
crossover models constant optimization response times etc.
Solve problems with multiple solutions;
available in MATLAB toolbox
Particle swarm Mimics bird and fish movement behavior Cannot work out the problems of non-
Speed of the researching is fast; calculation in [52–62],
optimization PSO is simple in comparison to other coordinate system; easily suffers from the [76]
methods; can be completed easily partial optimism etc.
Simulated Mimics an analogy between the way in Can deal with highly nonlinear models, There is a clear tradeoff between the quality [61–
annealing which a metal cools and freezes into a chaotic and noisy data and many constraints; of the solutions and the time required to 64,76]
minimum energy crystalline structure (the robust and general technique; flexible with compute; tailoring work is required to
annealing process) ability to approach global optimality; quite account for different classes of constraints
Fine-tuning of parameters of the algorithm
versatile as it does not rely on any restrictive
properties of the model can be rather delicate; significant effect
upon the quality of the outcome
Ant algorithms Inspired by the pheromone-based strategy of Algorithm has the strength in both local and Random initialization; algorithm has several [65–68]
ants foraging in nature; foraging behavior of global searches; implemented with several parameters; parameters need to be tuned;
ants is based on finding the shortest path optimization problems probabilistic approach in the local search
between source and their nests
Bee-inspired Based on the intelligent foraging behavior of Algorithm has local search and global search Random initialization; algorithm has several [69–81],
algorithms honey bee ability; implemented with several parameters
optimization problems; easy to use; available
for hybridization combination with other
algorithms
Harmony Based on improvisation process of jazz Does not require differential gradients, thus it Complex solving process [61–62],
search musicians can consider discontinuous functions as well [76,82–
as continuous functions; can handle discrete 83]
variables as well as continuous variables;
does not require initial value setting for the
variables; free from divergence; ability to
perform global and local search
Biogeography- Biogeography is the science of studying the Fast computation time; good convergence Poor in exploiting the solutions; no provision [77–81],
based behavior of species in nature against time accuracy for selecting the best members from each
optimization and space and species immigration and generation; sometimes many infeasible
(BBO) emigration between habitats, which is a solutions are generated
probable solution of the problem
Gravitational Based on Newton’s law of gravitation and the Good calculation accuracy; fast convergence Suffering from premature convergence [84–86]
search second law of motion speed problem sometime
algorithm GSA obtains
the optimal solution when particles attract
one another and bond
together to become a large entity
Imperialist The ICA based on a socio-politically inspired High convergence accuracy, appropriate for Complex process [87–89]
competition optimization strategy optimization of nonlinear hybrid power
algorithm generation system problems with high
dimensions
Tabu search Tabu search is a metaheuristic search Enhances the performance of local search; [61–
method employing local search methods fast computation 62,76,90–
used for mathematical optimization 91]
Hybrid Developed by using two or more algorithms Better accuracy in results; takes less Increased complexity; difficult to code [91–99]
optimization computational time (in some cases); much
techniques more competitive than any individual
method

to obtain the optimal number of photovoltaic panels, wind although it suffers from some shortcomings. Table 5 shows a
turbines and storages units ensuring the minimum cost and full summary of the discussed new generation optimization techniques.
availability of energy to meet load requirements. Authors con-
cludes that fuzzy-adaptive GA is better than standard GA after
comparing results of analysis using these two techniques. Mukh-
taruddin et al. [98] used a hybrid Iterative-Pareto-Fuzzy (IPF) 4. Discussion
technique to obtain the best compromised solution between PV–
wind–battery hybrid system that yields minimum cost and max- Concise sizing methodologies for PV–wind based hybrid sys-
imum reliability in a study location of Kuala Terengganu, Malaysia. tems including various requirements, criteria conditions and
Multi Criteria Decision Analysis (MCDA) optimization approach, a execution process with mathematical models of PV system, wind
procedure based on both Multi Objective Genetic Algorithm generator and battery bank are presented in this review. Various
(MOGA) and Multi Criteria Decision Making (MCDM) is proposed optimization techniques have been utilized by researchers to
by Alsayed et al. [99] for optimal design of grid connected design hybrid renewable energy systems which are also included
photovoltaic–wind hybrid system. in this study. Optimization problems of sustainable energy systems
Among all the discussed new generation algorithms in this section, become more and more complex, especially when more number of
GA is found to be mostly used in PV–wind hybrid system sizing renewable sources, are integrated together.
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 767

Table 6
Various techniques used in PV–wind based hybrid system studies during 1996–
2015.

Techniques Number of published studies Year

Graphic construction technique 2 1996


Iterative approach 7 1996–
2007
Probabilistic method 3 1996–
2006
Trade off approach 2 1992–1998
Linear Programming 6 1997–2014
Genetic algorithm 15 2006–
2015
Particle swarm optimization 11 2007–2015
Simulated annealing 4 2010–2015
Bee based algorithms 4 2011–2015
Fig. 5. Trend in the use of hybrid algorithms in PV–wind based hybrid system
Ant based algorithms 2 2006–
2013 research 2009–2015.
Bio geography based optimization 2 2013–2015
Gravitational search algorithm 1 2015
Tabu search 4 2014–2015 with other nature inspired techniques which are used in different
Harmony search 4 2014–2015 optimization studies, also seem promising to enrich the future
Imperialist competition algorithm 2 2015
research in renewable energy based hybrid energy system opti-
Hybrid algorithms 9 2009–
2015 mization studies [104,105].

In this study optimization methods are divided into two 5. Conclusion


categories e.g. traditional and new generation methods. Tradi-
tional approaches follows a rigorous procedure and also have This article presents an update literature review of sixteen
some drawbacks like rigid iterations, less flexibility, slow conver- types of optimization techniques including hybrid algorithms used
gence speed, needs more computation time, cannot deal with in PV–wind based hybrid energy system research and develop-
dynamic changing etc. On the other hand new generation ment along with PV–wind based hybrid system sizing methodol-
approaches are faster, flexible than traditional approaches with ogy. It is found that different optimization methods used have
good convergence speed and efficient global search solutions. Both different convergence speed, accuracy level; performance effi-
types of optimization methods are explained with their strengths- ciency and computation speed so selection of suitable approach
weaknesses in tabular form for better understanding. may change with user requirements, type of applications etc. None
The trend shows that new generation algorithms (evolution- of the individual methods could perform better than all the other
ary-heuristic–met heuristic) are much more widely accepted and methods on all kinds of problems. However, new generation
used in literature during last five years. The trend of number of optimization approaches like artificial intelligence based, heuristic
papers published using various techniques during period 1996– approaches are found to be more acceptable than traditional
2015 are shown in Table 6. approaches because of their ability to search local and global
Table 6 shows that old techniques like graphical construction optima, good calculation accuracy and fast convergence speed. The
method, iterative method, probabilistic method and trade off hybrid optimization techniques (using two or more optimization
approach were used almost a decade ago. Now-a-days these techniques together) are found best than single optimization
techniques are almost not used by researchers due to their draw- methods. The nature inspired optimization techniques and hybrid
backs. However, in recent years, researchers are mostly using optimization techniques will be important for further exploration
artificial intelligence (AI) based, nature inspired heuristic–met in future research to face complexity and challenges of PV–wind
heuristic algorithms like TS, GA, SA, HS, BBO, ACO, GCA, ICA etc. based hybrid systems.
for PV–wind based hybrid system optimization studies. Out of
these GA is mostly used AI technique than others.
For several complex problems a simple new generation algo- References
rithm might not be good enough to find the desired solution and
could fail to obtain a convenient solution. This clearly shows the [1] Sinha S, Chandel SS. Review of software tools for hybrid renewable energy
systems. Renewable Sustainable Energy Rev 2014;32:192–205.
need for hybridization of two or more algorithms efficiently. [2] Zhou W, Lou C, Li Z, Lu L, Yang H. Current status of research on optimum
Hybrid algorithms are found to be more reliable to use for better sizing of stand-alone hybrid solar–wind power generation systems. Appl
accuracy. The use of hybrid techniques results in improving the Energy 2010;87:380–9.
[3] Erdinc O, Uzunoglu M. Optimum design of hybrid renewable energy systems:
speed of convergence along with quality of the solutions obtained. overview of different approaches. Renewable Sustainable Energy Rev
During the period 2009–2015 various combinations of hybrid 2012;16:1412–25.
algorithms like SA-TS, GA-SVM, and SA-PSO etc. are being used [4] Fadaee M, MAM. Radzi. Multi-objective optimization of a stand-alone hybrid
renewable energy system by using evolutionary algorithms: a review.
in most of the studies. Although hybrid algorithms are more Renewable Sustainable Energy Rev 2012;16:3364–9.
complex than single algorithm yet the trend in the use of hybrid [5] Luna-Rubio R, Trejo-Peres M, Vargas-Vazquez D, Rı´os-Moreno GJ. Optimal
algorithms in PV–wind based hybrid system research is growing as sizing of renewable hybrids energy systems: a review of methodologies. Sol
Energy 2012;86:1077–88.
shown in Fig. 5. Further exploration of hybrid algorithms is an
[6] Khatib T, Mohamed A, Sopian K. A review of photovoltaic systems size
important future research area to face complexity and challenges optimization techniques. Renewable Sustainable Energy Rev 2013;22:454–65.
of hybrid system integration and design. [7] Upadhyay S, Sharma MP. A review on configurations, control and sizing
Some more algorithms like Cuckoo search [100], Bat algorithm methodologies of hybrid energy systems. Renewable Sustainable Energy Rev
2014;38:47–63.
[101], Firefly algorithm [102], Chaotic Ant swarm optimization [8] Chauhan A, Saini RP. A review on integrated renewable energy system based
[103], Cultural Algorithm [104], Memetic Algorithm [104] along power generation for stand-alone applications: configurations, storage
768 S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769

options, sizing methodologies and control. Renewable Sustainable Energy [40] Yang H, Zhou W, Lou C. Optimal design and techno-economic analysis of a
Rev 2014;38:99–120. hybrid solar–wind power generation system. Appl Energy 2009;86(2):163–9.
[9] Bourennani F, Rahnamayan S, Naterer GF. Optimal design methods for hybrid [41] Bilal BO, Sambou V, Ndiaye PA, Kébé CMF, Ndongo M. Optimal design of a
renewable energy systems. Int J Green Energy 2015;12:148–59. hybrid solar–wind–battery system using the minimization of the annualized
[10] Bhandari B, Lee KT, Lee GY, Cho YM, Ahn SH. Optimization of hybrid cost system and the minimization of the loss of power supply probability
renewable energy power systems: a review. Int J Precis Eng Manuf Green (LPSP). Renewable Energy 2010;35(10):2388–90.
Technol 2015;2(1):99–112. [42] AESAA Nafeh. Optimal economical sizing of a PV–wind hybrid energy system
[11] Kaldellis JK. Stand-alone and hybrid wind energy systems: technology, using genetic algorithm. Int J Green Energy 2011;8:25–43.
energy storage and applications. Woodhead Publishing Limited and CRC [43] Abdullrahman AAS, Addoweesh KE. Optimum sizing of hybrid PV/wind/
Press LLC; 2010. battery/diesel system considering wind turbine parameters using genetic
[12] Maghraby HAM, Shwehdi MH, Al-Bassam GK. Probabilistic assessment of algorithm. In: IEEE international conference on power and energy (PECon);
photovoltaic (PV) generation systems. IEEE Trans Power Syst 2002;17 2–5 December 2012. Malaysia, p. 121–26.
(1):205–8. [44] Atia R, Yamada N, Optimization of a PV–wind–diesel system using a hybrid
[13] Shrestha GB, Goel L. A study on optimal sizing of stand-alone photovoltaic genetic algorithm. In: IEEE electrical power and energy conference; 2012. p.
stations. IEEE Trans Energy Convers 1998;13(4):373–8. 80–85.
[14] Al-Ashwal AM, Moghram IS. Proportion assessment of combined PV–wind [45] Merei G, Berger C, Sauer DW. Optimization of an off-grid hybrid PV–wind–
generating systems. Renewable Energy 1997;10(1):43–51. diesel system with different battery technologies using genetic algorithm.
[15] Borowy BS, Salameh ZM. Methodology for optimally sizing the combination Sol Energy 2013;97:460–73.
of a battery bank and PV array in a wind/PV hybrid system. IEEE Trans Energy [46] Abbes D, Martinez A, Champenois G. Life cycle cost, embodied energy and
Convers 1996;11(2):367–73. loss of power supply probability for the optimal design of hybrid power
[16] Markvart T. Sizing of hybrid PV–wind energy systems. Sol Energy 1996;59 systems. Math Comput Simul 2014;98:46–62.
(4):277–81. [47] Shi JH, Zhu HJ, Cao GY. Design and techno-economical optimization for
[17] Yang HX, Lu L, Zhou W. A novel optimization sizing model for hybrid solar– stand-alone hybrid power systems with multi-objective evolutionary algo-
wind power generation system. Sol Energy 2007;81(1):76–84. rithms. Int J Energy Res 2007;31:315–28.
[18] Kellogg W, Nehrir MH, Venkataramanan G, Gerez V. Generation unit sizing [48] Mostofi F, Shayeghi H. Feasibility and optimal reliable design of renewable
and cost analysis for stand-alone wind, photovoltaic and hybrid wind/PV hybrid energy system for rural electrification in Iran. Int J Renewable Energy
systems. IEEE Trans Energy Convers 1998;13(1):70–5. Res 2012;2(4):574–82.
[19] Kellogg W, Nehrir MH, Venkataramanan G, Gerez V. Optimal unit sizing for a [49] Shadmand MB, Balog RS. Multi-objective optimization and design of photo-
hybrid wind/photovoltaic generating system. Electr Power Syst Res voltaic–wind hybrid system for community smart DC micro grid. IEEE Trans
1996;39:35–8. Smart Grid 2014;5(5):2635–43.
[20] Diaf S, Diaf D, Belhamel M, Haddadi M, Louche A. A methodology for optimal [50] Tégani I, Aboubou A, Ayad MY, Becherif M, Saadi R, Kraa O. Optimal sizing
sizing of autonomous hybrid PV/wind system. Energy Policy 2007;35:5708–18. design and energy management of stand-alone photovoltaic/wind generator
[21] Prasad AR, Natarajan E. Optimization of integrated photovoltaic–wind power systems. Energy Procedia 2014;50:163–70.
systems with battery storage. Energy 2006;31:1943–54. [51] Bilal BO, Nourou D, Kébé CMF, Sambou V, Ndiaye PA, Ndongo M. Multi-
[22] Borowy BS, Salameh ZM. Methodology for optimally sizing the combination objective optimization of hybrid PV/wind/diesel/battery systems for decen-
of a battery bank and PV array in a wind/PV hybrid system. IEEE Trans tralized application by minimizing the levelized cost of energy and the CO2
Energy Convers 1996;11(2):367–75. emissions. Int J Phys Sci 2015;10(5):192–203.
[23] Ashok S. Optimised model for community-based hybrid energy system. [52] Kennedy J, Eberhart R. Particle swarm optimization. In: Proceedings of IEEE,
Renewable Energy 2007;32(7):1155–64. international conference on neural networks, 1995; 4(2):1942–48.
[24] Bagul AD, Salameh ZM, Borowy B. Sizing of a stand-alone hybrid wind [53] Poli R, Kennedy J, Blackwell T. Particle swarm optimization. Swarm Intell
photovoltaic system using a three-event probability density approximation. 2007;1(1):33–57.
Sol Energy 1996;56(4):323–35. [54] Bashir M, Sadeh J. Size optimization of new hybrid stand-alone renewable
[25] Karaki SH, Chedid RB, Ramadan R. Probabilistic performance assessment of energy system considering a reliability index. In: Eleventh IEEE international
autonomous solar–wind energy conversion systems. IEEE Trans Energy conference in environment and electrical engineering (EEEIC); 2012. p. 989–94.
Convers 1999;14:766–72. [55] Lee TY, Chen CL. Wind–photovoltaic capacity coordination for a time-of-use
[26] Tina G, Gagliano S, Raiti S. Hybrid solar/wind power system probabilistic rate industrial user. IEEE Trans Renewable Power Gener 2009;3(2):152–67.
modeling for long-term performance assessment. Sol Energy 2006;80 [56] Kaviani AK, Riahy GH, SHM. Kouhsari. Optimal design of a reliable hydrogen
(5):578–88. based stand-alone wind/PV generating system, considering component
[27] Chedid R, Akiki H, Rahman. S. A decision support technique for the design of outages. Renewable Energy 2009;34(11):2380–90.
hybrid solar–wind power systems. IEEE Trans Energy Convers [57] Bansal AK, Gupta RA, Kumar R. Optimization of hybrid PV/wind energy
1998;13:76–82. system using Meta Particle Swarm Optimization (MPSO). In: IEEE 2011 India
[28] Gavanidou ES, Bakirtzis AG. Design of a standalone system with renewable international conference on power electronics (IICPE), 2010. p. 1–7.
energy sources using trade off methods. IEEE Trans Energy Convers [58] Sharafi M, ELMekkawy TY. Multi-objective optimal design of hybrid renew-
1992;7:42–8. able energy systems using PSO-simulation based approach. Renewable
[29] Chedid R, Rahman S. Unit sizing and control of hybrid wind–solar power Energy 2014;68:67–79.
systems. IEEE Trans Energy Convers 1997;12(1):79–85. [59] Pirhaghshenasvali M, Asaei B. Optimal modeling and sizing of a practical
[30] Huneke F, Henkel J, González JAB, Erdmann G. Optimization of hybrid off- hybrid wind/PV/diesel generation system. In: The fifth power electronics,
grid energy systems by linear programming. Energy Sustainability Soc drive systems and technologies conference (PEDSTC 2014), Tehran, Iran;
2012;2(7):1–19. 2014. p. 506–11.
[31] Nogueira CEC, Vidotto ML, Niedzialkoski RK, Melegari de Souza SN, Chaves LI, [60] Borhanazad H, Mekhilef S, Ganapathy VG, Delshad MM, Mirtaheri A.
Edwiges T, et al. Sizing and simulation of a photovoltaic–wind energy Optimization of micro-grid system using MOPSO. Renewable Energy
system using batteries, applied for a small rural property located in the south 2014;71:295–306.
of Brazil. Renewable Sustainable Energy Rev 2014;29:151–7. [61] Maleki A, Askarzadeh A. Comparative study of artificial intelligence techni-
[32] Lee JY, Chen CL, Chen HC. A mathematical technique for hybrid power ques for sizing of a hydrogen-based stand-alone photovoltaic/wind hybrid
system design with energy loss considerations. Energy Convers Manage system. Int J Hydrog Energy 2014;39:9973–84.
2014;82:301–7. [62] Maleki A, Ameri M, Keynia F. Scrutiny of multifarious particle swarm
[33] Saif A, Gad Elrab K, Zeineldin HH, Kennedy S, Kirtley JL. Multi-objective optimization for finding the optimal size of a PV/wind/battery hybrid
capacity planning of a PV–wind–diesel–battery hybrid power system. In: system. Renewable Energy 2015;80:552–63.
IEEE international conferences; 2010. [63] Kirkpatrick S. Optimization by simulated annealing: quantitative studies.
[34] Nagabhushana AC, Jyoti R, Raju AB. Economic analysis and comparison of J Stat Phys 1984;34(5–6):975–86.
proposed hres for stand-alone applications at various places in Karnataka [64] Ekren O, Ekren BY. Size optimization of a PV/wind hybrid energy conversion
state. IEEE PES Innovative Smart Grid Technol—India 2011:380–5. system with battery storage using simulated annealing. Appl Energy
[35] Deb Kalyanmoy. Multi-objective optimization using evolutionary algorithms. 2010;87:592–8.
vol. 2012. Chichester: John Wiley & Sons; 2001. [65] Dorigo M, Gambardella LM. Ant colony system: a cooperative learning
[36] Goldberg David Edward. Genetic algorithms in search, optimization, and approach to the traveling salesman problem. In: IEEE transactions on
machine learning. . Reading Menlo Park: Addison-Wesley; 1989. evolutionary computation 199; 1.1(7):53–66.
[37] Koutroulis E, Kolokotsa D, Potirakis A, Kalaitzakis K. Methodology for optimal [66] Dorigo M, Birattari M, Stutzle T. Ant colony optimization. IEEE Comput Intell
sizing of stand-alone photovoltaic/wind-generator systems using genetic Mag 2006;1.4:28–39.
algorithms. Sol Energy 2006;80(9):1072–88. [67] Menshsari A, Ghiamy M, Mousavi MMM, Bagal HA. Optimal design of hybrid
[38] Koutroulis E, Kolokotsa D. Design optimization of desalination systems water–wind–solar system based on hydrogen storage and evaluation of
power-supplied by PV and W/G energy sources. Desalination 2010;258(1– reliability index of system using ant colony algorithm. Int Res J Appl Basic Sci
3):171–81. 2013;4(10):3582–600.
[39] Yang H, Zhou W, Lu L, Fang Z. Optimal sizing method for stand-alone hybrid [68] Xu D, Kang L, Cao B. Graph-based ant system for optimal sizing of standalone
solar–wind system with LPSP technology by using genetic algorithm. Sol hybrid wind/PV power systems. Computational inelligence. Berlin Heidel-
Energy 2008;82(4):354–67. berg: Springer; 2006. p. 1136–46.
S. Sinha, S.S. Chandel / Renewable and Sustainable Energy Reviews 50 (2015) 755–769 769

[69] Passino KM. Biomimicry of bacterial foraging for distributed optimization [87] Atashpaz-Gargari E, Lucas C. Imperialist competitive algorithm: an algorithm
and control. IEEE Control Syst Mag 2002;22(3):52–67. for optimization inspired by imperialistic competition. In: IEEE congress on
[70] Bazyar R, Valipoor KH, Javadi MR, Valizade M, Kord H. Optimal design and evolutionary computation, Singapore, 2007; p. 4661–4667.
energy management of stand-alone wind/PV/diesel/battery using bacterial [88] Gharavi H, Ardehali MM, Ghanbari-Tichi S. Imperial competitive algorithm
foraging algorithm. In: Eighth national congress; 2010, p. 1–14. optimization of fuzzy multi-objective design of a hybrid green power system
[71] Karaboga D, Basturk B. A powerful and efficient algorithm for numerical with considerations for economics, reliability, and environmental emissions.
function optimization: artificial bee colony (ABC) algorithm. J Global Optim Renewable Energy 2015;78:427–37.
2007;39:459–71. [89] Ranjbar MR, Kouhi S. Sources response for supplying energy of a residential
[72] Karaboga D, Basturk B. Artificial bee colony (ABC) optimization algorithm for load in the form of on-grid hybrid systems. Electr Power Energy Syst
solving constrained optimization problems. Foundations of fuzzy logic and 2015;64:635–45.
[90] Glover F, Laguna M. Tabu search. US: Springer; 1999.
soft computing. Berlin Heidelberg: Springer; 2007. p. 789–98.
[91] Katsigiannis YA, Georgilakis PS, Karapidakis ES. Hybrid simulated annealing
[73] Nasiraghdam H, Jadid S. Optimal hybrid PV/WT/FC sizing and distribution
—Tabu search method for optimal sizing of autonomous power systems with
system reconfiguration using multi-objective artificial bee colony (MOABC)
renewables. IEEE Trans Sustainable Energy 2012;3(3):330–8.
algorithm. Sol Energy 2012;86:3057–71.
[92] Askarzadeh A. A discrete chaotic harmony search-based simulated annealing
[74] Maleki A, Askarzadeh A. Artificial bee swarm optimization for optimum
algorithm for optimum design of PV/wind hybrid system. Sol Energy
sizing of a stand-alone PV/WT/FC hybrid system considering LPSP concept. 2013;97:93–101.
Sol Energy 2014;107:227–35. [93] Tutkun N. Minimization of operational cost for an off-grid renewable hybrid
[75] Tudu B, Majumder S, Mandal KK, Chakraborty N. Optimal unit sizing of system to generate electricity in residential buildings through the SVM and
stand-alone renewable hybrid energy system using bees algorithm. In: IEEE the BCGA methods. Energy Build 2014;76:470–5.
international conference on energy, automation, and signal (ICEAS), 2011; [94] Khatib T, Mohamed A, Sopian K. Optimization of a PV/wind micro-grid for
pp. 1–6. rural housing electrification using a hybrid iterative/genetic algorithm: case
[76] Maleki A, Pourfayaz F. Optimal sizing of autonomous hybrid photovoltaic/ study of Kuala Terengganu, Malaysia. Energy Build 2012;47:321–31.
wind/battery power system with LPSP technology by using evolutionary [95] Dehghan S, Kiani B, Kazemi A, Parizad A. Optimal sizing of a hybrid wind/PV
algorithms. Sol Energy 2015;115:471–83. plant considering reliability indices. World Acad Sci Eng Technol 2009;56
[77] Simon D. Biogeography-based optimization. IEEE Trans Evol Comput (32):527–35.
2008;12(6):702–13. [96] Zhou T, Sun W. Optimization of battery–super capacitor hybrid energy
[78] Kumar R, Gupta RA, Bansal AK. Economic analysis and power management of storage station in wind/solar generation system. IEEE Trans Sustainable
a stand-alone wind/photovoltaic hybrid energy system using biogeography Energy 2014;5(2):408–15.
based optimization algorithm. Swarm Evol Comput 2013;8:33–43. [97] Abdelhak BJ, Najib E, Abdelaziz H, Hnaien F. Optimum sizing of hybrid PV/
[79] Liang JJ, Qin AK, Suganthan PN, Baskar S. Comprehensive learning particle wind/battery using fuzzy-adaptive genetic algorithm in real and average
swarm optimizer for global optimization of multimodal functions. IEEE Trans battery service life. In: IEEE international symposium on power electronics,
Evol Comput 2006;10:281–95. electrical drives, automation and motion; 2014. , p.871–876.
[80] Mallipeddi R, Suganthan PN, Pan QK, Tasgetiren MF. Differential evolution [98] Mukhtaruddin RNSR Rahman HA, Hassan MY, Jamian JJ. Optimal hybrid
algorithm with ensemble of parameters and mutation strategies. Appl Soft renewable energy design in autonomous system using Iterative-Pareto-
Fuzzy technique. Electr Power Energy Syst 2015;64:242–9.
Comput 2010;11:1679–96.
[99] Alsayed M, Cacciato M, Scarcella G, Scelba G. Design of hybrid power
[81] Gupta RA, Kumar R, Bansal AK. BBO-based small autonomous hybrid power
generation systems based on multi criteria decision analysis. Sol Energy
system optimization incorporating wind speed and solar radiation forecast-
2014;105:548–60.
ing. Renewable Sustainable Energy Rev 2015;41:1366–75.
[100] Yang XS, Deb S. Engineering optimization by cuckoo search. Int J Math Model
[82] Geem ZW, Kim JH, Loganathan GV. A new heuristic optimization algorithm:
Numer Optim 2010;1(4):330–43.
harmony search. Simulation 2001;76:60–8. [101] Yang XS. Bat algorithm for multi-objective optimization. Int J Bio-Inspired
[83] Maleki A, Askarzadeh A. Optimal sizing of a PV–wind–diesel system with Comput 2011;3:267–74.
battery storage for electrification to an off-grid remote region: a case study [102] Łukasik S, Żak S. Firefly algorithm for continuous constrained optimization
of Rafsanjan, Iran. Sustainable Energy Technol Assess 2014;7:147–53. tasks. Computational collective intelligence. Semantic web, social networks
[84] Rashedi E, Nezamabadi-Pour H, Saryazdi S. GSA: a gravitational search and multiagent systems. Springer-Berlin-Heidelberg; 2009. p. 97–106.
algorithm. Inf Sci 2009;179(13):2232–48. [103] Li Y, Wen Q, Li L, Peng H. Hybrid chaotic ant swarm optimization. Chaos,
[85] Rashedi E, Nezamabadi-Pour H, Saryazdi S. BGSA: binary gravitational search Solitons Fractals 2009;42(2):880–9.
algorithm. Nat Comput 2010;9:727–45. [104] Brownlee J. Clever algorithms: nature-inspired programming recipes. 1st ed..
[86] Wu K, Zhou H, An S, Huang T. Optimal coordinate operation control for LuLu: Jason Brownlee; 2011.
wind–photovoltaic–battery storage power-generation units. Energy Convers [105] Yang Xin-She. Nature-inspired metaheuristic algorithms. 2nd ed. UK: Luniver
Manage 2015;90:466–75. Press; 2010.