A Comprehensive review on Establishment of

Isolated Microgrid with Renewable Sources for

Remote Islands
M.C.Lavanya, PhD Scholar Dr.Shasidhar Sampathrao., Asst. Professor
School of Electrical Sciences School of Electrical Sciences
Indian Institute of Technology Indian Institute of Technology
Goa, India Goa, India
lavanya19242202@iitgoa.ac.in ssd@iitgoa.ac.in

Abstract—Daily increase of energy demands and utilization of power. Finding an optimal sizing, efficient control strategies
the resources, hikes every year by year. Energy extraction from along with the functionality prediction of these RESs,
renewable distributive sources to support various load demands is conventional sources and ESS for a given application is
tremendously increasing. 50 % of the rural population across the essential for the reliable, efficient and economical operation
globe at remote areas are still struggling for basic energy needs. of a microgrid.
Hybrid Microgrids (MGs) designed with Renewable energy
sources (RESs) could serve as an optimal sustainable solution to II. ESTABLISHMENT OF HYBRID MICROGRID WITH
support these isolated communities. As these remote Islands/ RENEWABLE SOURCES
areas are not easily accessible to the mainland, uninterrupted
energy supply with proper demand side management need to be Microgrids are small networks that work as a cluster of
executed with real-time controls. The proposed Hybrid MG loads, storage units and micro-sources operating as a single
system comprises of RESs, Diesel Gensets and Energy Storage controllable network that produces and distribute power in
systems (ESS). A comprehensive review has made to work on remote areas such as villages, industrial complexes or even in
Hybrid RE power prediction algorithm and its integration with remote islands
energy management system (EMS) for designing consistent,
reliable and scalable Isolated Hybrid Microgrid for a remote The reliability of the Microgrid is improved by optimal
island without instituting negative impacts on the local sizing of the sources and integrating the control system with
environment. appropriate RE power prediction algorithm. Communications
systems need to be equipped for systems established at
remotes areas for real time monitoring and control, bidding
to satisfy the energy needs of the isolated area population.
Keywords—Hybrid Microgrid, Renewable Energy sources,
Energy Management System, Power prediction and Remote
Islands.

I. INTRODUCTION
The role of energy generation is one of the most
important factors for the development of any country. The
main sources of energy are now depleting and conventional
systems have failed to provide reliable electricity in rural
areas or remote areas which are away from the grid. Hence
now it is essential to look towards alternative energy sources
like renewable sources PV, wind, biomass, hydro etc. A
hybrid energy system is defined as the component
combination of two or more types of power generation Fig 1. Basic structure of proposed MG System
system such as solar energy system integrated with wind
turbine system to form a hybrid renewable energy system. Microgrid operates either in isolation or grid connected
Since the power output of these renewable energy is mode based on the type of the source they manage, control,
ultimately depending on climatic conditions such as operate and application, Microgrid can be classified as given
temperature, solar irradiance, wind speed and etc., the in the flowchart.
instability of the system output is compensated by adding a
suitable energy storage system to the hybrid energy system.
The International Energy Agency (IEA) anticipates that more
than 50% of the rural population currently without energy
access where these systems can serve as they are much
appropriate to use in remote areas where utility grid is
inaccessible.
Microgrids paves as a platform to integrate distributed
generation such as renewable energy sources and
conventional sources in both grid-connected and isolated
power systems thereby achieving reliable supply of electrical Fig 2. Flowchart of MG classification
 Consistent demand side management in a Microgrid inverters/ converters equipped, the control on voltage and
(MG) is essential as it depends on distributed sources such as frequency will be employed for seamless power transfer from
photovoltaic and wind turbines which are highly intermittent the MG to the Utility/ Load. Hence optimal sizing and proper
to predict and their generation may fluctuate significantly co-ordination of controls need to be executed for
depending on the availability of the primary sources. The uninterrupted stable MG operation [2].
supply-demand balancing problem becomes even more
important when the microgrid is operating in stand-alone III. MATHEMATICAL MODELLING OF THE
mode where only limited supply is available to balance the MICROGRID SUBSYSTEMS
demand. Hence an efficient Energy Management Control The standalone Microgrid configuration of this study
System [EMCS] integrated with RE power forecasting tool contains PV systems, and wind turbine as renewable
need to be employed for proper uninterrupted operation of generation sources. Whereas, diesel generators as the
MG leading to a sustainable solution to the energy demand. conventional generation source and the battery storage
Control Strategies in Microgrid system system (Lithium-ion battery) is used to store the surplus of
power and discharge when there is a deficit. [3]
The Consortium for Electric Reliability Technology
Solutions (CERTS) defines “Microgrid as an aggregation of MODELING OF WIND ENERGY GENERATOR
loads and micro source providing both power and heat. It A wind turbine converts kinetic energy from the wind
operates as a single system which could separate or island into electricity. The blades of a wind turbine turn between 13
from the bulk power system if problems arise on the grid, and 20 revolutions per minute, depending on their
and reconnects to the grid once they are resolved”. [1] technology, at a constant or variable velocity. Depending on
The Scope of microgrid control are as follows: the speed of the wind the amount of electrical ac energy will
be produced. It has be assessed that wind energy has lowest
relative greenhouse gas emissions, least water consumption
 New distributive sources can be added to the MG without demands and the most favorable social impacts compared to
any complex modifications. photovoltaic, hydro, geothermal, coal and gas.[4] Basically
there are two types of wind generators i.e., vertical axis and
horizontal axis wind turbine. They can be used to generate
 MG can be automated and operated from remote in Real- electricity both onshore and offshore. Wind Turbines can be
time. combined to form clusters called “wind farms”, which plays
a prominent role in supporting peak demands of the grid.
 MG can be operated both in isolated and grid connected Apart from generating electricity they can also be used for
grain-grinding, water pumping, charging batteries. Power
mode depending on the energy demand
generation Pw of the wind turbine can be calculated with the
equation.
 Power quality issues can be recovered and suitable power
control techniques can be augmented. C p. ρ . A V w3
Pw =
2
 Consistent demand side management can be employed Where C p is the power coefficient, Vw is the wind
with proper energy management control system of MG speed. A is the area of swept by the rotor and ρ is the air
density. Amount of aerodynamic torque Tw is given by
Pw
The Distributive Generation (DG) from RESs T w=
(Renewable Energy Sources) and conventional sources can Ww
be categorized as dc sources (fuel cells, photovoltaic cells
and battery storage) and ac sources (Wind Turbine and Where Pw is the power extracted from the wind and Ww
Diesel Gensets). is the turbine rotor speed.
All the combined generated power fed to the Utility/
Load through power electronics devices such as Converters/ MODELING OF SOLAR PHOTOVOLTAIC (SPV)
Inverters. Hence control of these power electronic (PE) GENERATION SYSTEM
devices plays a vital role in power conversion, stability and
consistent power supply. Depending on the control strategies A photovoltaic Generation system uses one or more solar
of PE devices, the MG control can be classified as in the panels to convert solar energy to electrical energy. The basic
flowchart. unit of the PV system is photovoltaic cell, These PV modules
are connected in series or parallel to form a PV array. The
amount of electrical dc energy produced by the PV panels
depends on its position, ambient temperature and the
intensity of solar irradiation perceived by the panel. The
maximum energy acquisition can be obtained by tracking the
Sun’s path and positioning the PV array, the following
equations provide the power generated with and without
tracking.
Fig 3. Flowchart of MG control classification
Without Tracking
The system layout/ sizing and the type of sources will
detail the PE devices implementation, depending on the V busDC =V npanel∗N S
 Where V busDC is the dc bus voltage, V n panel is the nominal For extremely small loads (a few lights), one or two no. of
voltage of the PV panel, Ns is the number of panels in series SPV panels, say 50 W make economic sense. PV system will
and Np is the number of panels in parallel. be easier to work with and less expensive than SWT in this
case. Wind – solar hybrid systems become attractive as loads
I sc∗G∗V npanel∗N s∗N p increase above a few hundred watts. A hybrid wind solar
P= requires sizing and siting the PV array for summer
LF conditions.
Where I sc is the short circuit current of panel, G is the At low wind speed sites, small wind turbines with relatively
irradiance on the surface of the panels in kW/m 2 and L.F is large diameter rotors relative to their generator capacity will
the loss factor. In this case output power practically does not keep a continuous charge flowing to the battery rather than
depend on ambient temperature. turbines with small rotors and high power ratings.
With Maximum Power Point Tracking In order to determine the inverter size, the inverter
continuous and surge rating will help in sizing the inverter.
Ct

P=
Pn∗G 1+
( 100 ( T c −25 ) )
∗N s ¿ N p As a rule of thumb, the inverter should be sized at 125 % of
expected load. Besides this, the inverter should have
disconnected switches on both ac and dc sides and should
FS offer power factor correction.
Where Pn is the nominal power (Wp) of PV panel, T c is
the internal cell temperature, C t is the maximum power Charging state of batteries need to be considered, a battery
temperature coefficient and FS is the safety factor. need not to be drained to 20 % of their full charge and cannot
be charged up to 90 % of its full charge. Lead-acid batteries
can be charged up to 80% only
MODELING OF BATTERY SYSTEM
System sizing
In the absence of renewable sources like solar and wind,
This involves determination of the right practical size of the
battery system plays a vital role in ensuring uninterrupted
hybrid components so as to minimize the cost. Different
power in supplying loads. The capacity of the battery is
optimization techniques can be used for arriving at the
determine based on daily energy consumption (EDL) and the
optimum size. An oversizing of components may lead to an
time during which load is supply from the battery bank. The
increased system cost and under sizing will yield an
battery storage capacity (C wh) is calculated using the unreliable system.
equation
Different software tools such as HOMER, HYBRID 2 are
E DL . AD available for optimizing hybrid systems. The conventional
C wh= approach to designing a hybrid system is using the wind and
(η ¿ ¿ BDI . ηB . DOD)¿ solar meteorological data to arrive at a system configuration
Where AD is number of autonomy days (in the absence which will ensure an energy balance. The iterative
of renewables), DOD is the allowable depth of discharge of optimization will involve considering various combinations
battery, η BDI is the bi-directional inverter efficiency and η B of system component sizes, namely SPV system, wind
is the battery efficiency. system and battery sizes and arrive at the optimum
configuration giving a reliable supply at lowest Levelized
Optimal system modeling and sizing depends on the Cost of energy.
availability of the potential resources at the site and
following important factors need to be taken into account IV. CASE STUDY : LAKSHADWEEP ISLANDS -
before proceeding into design and implementation. KAVARATTI , U.T

Factors to be considered for installation of wind

Population of the area: 11000 (As per 2011 population census)
solar hybrid systems As mentioned, before proceeding into design and
implementation of RE generation, site selection and RE
The objective of installing a wind solar hybrid system is to potential estimation is majorly considered. The proposed
provide a reliable power supply; hence an optimum system is planned to design for a remote island Kavaratti,
configuration should be chosen so as to get maximum India, where minimal population exists and completely
benefits from the systems. Important factors [5] that need to depend on Diesel Gensets for their day to day energy
be considered are: demands. Installation of Autonomous Isolated-MG will be an
Site selection ideal sustainable solution for the island dweller’s energy
demands.
Resource assessment and commercial viability analysis need
to be carried out for verifying the energy potential at the site. Site Details
Metrological data can also be reviewed with nearby weather
stations. Best assessment yields return on investment within
certain period.
System configuration
System configuration depends on factors such as availability
of grid / alternate sources, critical loads to be supplied and
also the cost economics of the system.
  The 11 kV underground feeders mainly run along the
periphery of the island.

Renewable Energy potential estimation of Kavaratti

Wind Resource assessment

 Wind speed profile

Time series- wind speed profile: Plotted based on

(01.01.2018 – 01.01.2019) wind speed MERRA data in m/s
at an altitude of 50m.

Time series wind Speed_50m [m/s]

16
12
8
4
Fig 4. Map of Kavaratti, Lakshadweep Islands (UT) 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0
:00 :00 :00 :00 :00 :00 :00 :00 :00 :00 :00 :00 :00 :00 :00
17 05 17 05 17 05 17 05 17 05 17 05 17 05 17
/ 01 /27 /21 /19 /13 /09 /03 /29 /24 /19 /13 /09 /03 /29 /24
01 01 02 03 04 05 06 06 07 08 09 10 11 11 12
18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/ 18/
20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

Fig 6. Time series wind speed profile

 Average wind speed profile

Fig 5. Site details of Kavaratti Average Annual Wind Speed: 5.25 m/s (Based on 10 years
of MERRA data)
Kavaratti island has most of the Administrative offices of
Lakshadweep authority. The island has got its own
independent power supply arrangement with Diesel- Monthly Average Wind Speed profile
Generator Sets, which is not connected with main land or
10
other islands. 8.99
9
7.89
The local loads are catered through the diesel generating 8 7.48
Wind Speed (m/s)

system of 2280kW installed capacity, solar power plants of 7 6.03

1065 kW aggregate capacity and Bio mass gasifier of 250 6
4.93
kW capacity. Power supply at Kavaratti Island is made 5 4.26 4.27
available 24 hrs a day. During daytime 4 DG sets are under 3.78 3.49 3.69 4.05 3.7
4
operation and 5 sets are operated during evening hours. 3
Transmission and distribution 2
1
The electrical requirements of the Island are being met 0
primarily by the power generated from diesel generators. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
There is no existing power connectivity between the Fig 7. Monthly wind speed profile
mainland and the Islands of Lakshadweep.
 Wind Rose (Wind Direction)
 Transmission and distribution of power generated by
Power House is done by HT and LT lines.
Predominant Direction: West (W) & North of North East
 The 11kV HT overhead line is laid for 7kms. LT lines
(NNE)
constituting overhead lines and underground cables with
3-phases laid for around 22 km and single phase laid for
6 km.
 There are 3 Nos.11 kV underground feeders emerging
from Power House, viz., North, South & East feeders.
 These are feeding 17 Nos. of 11 kV/440 V Distribution
transformers scattered over various parts of island with
capacity ranging between 25 KVA to 160 KVA, making
an aggregate capacity of 1471 KVA.
 Fig 11. Normalized Power curve
20 kW Wind turbine power curve generated based on
01.01.2018 – 01.01.2019 wind speed MERRA data

Fig 8. Wind Direction profile

 Diurnal Profile

Hourly wind speed profile: Hour of peak Speed is 1.00

pm.

Fig 12. Estimated Wind Turbine Power curve

Wind power extracted from the estimated Wind turbine

power curve for the Wind speed data: Rated power, PR = 20
kW, Cut-in Wind speed, UC = 3 m/s, Rated Wind speed UR=
12 m/s, Cut-off Wind Speed, UF = 23 m/s
Fig 9. Hourly wind speed profile

 Temperature profile
Temperature profile: Average Temperature- 27.9°C
(Based on 01.01.2018 – 01.01.2019 MERRA data)

Fig 10. Temperature profile

 Wind power potential estimation

Normalized power curve for 20kW rated power -Wind

turbine

Fig 13. Wind power Extraction in kW

RE potential estimation based on MERRA data
Wind Potential Estimation
 300 Azad, H. B., [7] worked on statistical and the neural
network-based approaches to predict hourly wind speed data
250 of the subsequent year. The novelty of this study is to
forecast the general trend of the incoming year by designing
200 a data fusion algorithm through several neural networks.
150 Bali, V, [8] In this survey paper, the authors have done
detailed study of various researchers’ work. Suggested that
100 deep learning methods with LSTM (long short term memory)
can result in the better prediction of wind speed for power
50 generation.
0 Cadenas, E., [9] wind speed forecasting Hybrid models
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec consisting of Autoregressive Integrated Moving Average
Fig 14. Wind energy potential in MWh for Kavaratti (ARIMA) models and Artificial Neural Network (ANN)
140 models were developed. Statistical error measures such as the
128 124 mean error (ME), the mean square error (MSE) and the mean
120 Potential
Solar 118 Estimation in MW
100105 102 101 96 absolute error (MAE) were calculated to compare the three
97 96
87 methods.
80 80 77
60 Guo, Z.., [10] proposed a modified EMD-FNN (Empirical
Mode Decomposition based feed forward neural network for
40
Wind Speed Forecasting. This method is for decomposing
20 the nonlinear and non-stationery wind speed series into finite
0 and intrinsic mode functions(IMD) by using Hilbert Huang
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Transform. The Developed model is an efficient tool for
operation, planning and dispatching of Wind farms.

Fig 15.Solar photo voltaic energy potential in MWh of Lei, M. [11] reviewed the forecasting models above have
Kavaratti their own characteristics, and they can perform well in
different situations. NWP models are good at predicting
Inference large-scale area wind speed and can achieve better results in
long-term forecasting. Often they were used as input of time-
As the selected site is almost coming under offshore category
series models as ARMA, ANN, etc., and help them to obtain
and the wind speed data itself infers that the island packed
better results.
with high RE potential at the same time the RE sources are
highly intermittent in nature. Therefore, more accurate RE Mohandes, M. A., [12] introduces support vector machines
power prediction need to be employed and implemented with (SVM), the latest neural network algorithm, to wind speed
the EMCS for optimal RE exaction for the MG operation. prediction and compares their performance with the
multilayer perceptron (MLP) neural networks.
Renewable Sources and the necessity of its prediction
Guo, Z., [13] This paper proposes a new hybrid model for
RE power forecasting is an important foundation and
long-term wind speed forecasting based on the first definite
prerequisite for the prediction of RE power generation.
season index method and the Autoregressive Moving
Algorithms like Artificial neural networks (ANN) works on
Average (ARMA) models or the Generalized Autoregressive
very simplex principle and it models linear and nonlinear
Conditional Heteroskedasticity (GARCH) forecasting
systems. It predicts the output for an unseen input by
models. The forecasting errors are analyzed and compared
repeating the training patterns. ANN finds number
with the ones obtained from the ARMA, GARCH model, and
applications in the area of load forecasting, security
Support Vector Machine SVM; the simulation process and
assessment, fault location/diagnosis, solar radiation
results show that the developed method is simple and quite
estimation and prediction of wind speed etc. Short term
efficient for daily average wind speed forecasting
updates are less vital to some sites than accuracy and
significant time to onset, whereas effective long-term Barbounis, T.G. [14] Three local recurrent neural networks
forecasting requires not just a trigger of current events but are employed in this paper, providing 72 time-steps ahead
also deep knowledge on historical patterns of Seasonal winds forecasts of the wind speed and power at the Rokas’ wind
and site specific meteorological parameters. park on the Greek island of Crete. And also details that the
recurrent forecast models outperform the static rivals in
Therefore, Hybrid models combining statistical and neural
terms of forecast errors and the improvement gained over the
network models of prediction need to be employed for
persistent method.
medium and long-term forecasting to increase the possibility
of site- specific accurate prediction of RESs. Nair, K. R., [15] used models such as Artificial Neural
Network (ANN) method, Auto Regressive Integrated
Literature Review on RE prediction
Moving Average Method (ARIMA) and Hybrid method,
Ambach [6] focused on accurate short- and medium-term which is the combination of ARIMA and ANN and the
wind speed and power predictions. Statistical forecasting compared the results. The author concluded that if the
approaches is examined in detail. Subsequently, double cross behaviour of the wind speed is linear or non-linear, the
arrangement models are surveyed, a univariate seasonal hybrid model will give a better result than ANN model as
ARFIM-AAPARCH model and a multivariate periodic well as ARIMA model.
seasonal VAR-TARCH model. These models are
Shukur, O. B.., [16] In this paper, Artificial Neural
additionally used to give wind power estimates.
Network(ANN) and Kalman filter(KF) is combined into one
hybrid model to improve the accuracy in wind speed based on general algebraic modeling system. And also details
forecasting. In this two wind speed data sets were used from the advantages on programming the prediction of RESs,
meteorological environment. The advantages of the hybrid Genset operations and controllability on loads
model were the result of hybridizing the linear and nonlinear
Taha and Yasser [18] proposed a robust predictive control
parts of the ARIMA model and combing it with a KF to
handle stochastic uncertainty and ANN to handle the algorithm for an isolated MG. The EMS of the MG is
optimized with MILP results in operational cost reduction,
nonlinearity within hybrid KF-ANN model.
energy consumption, and gas emission due to diesel
Inference generation.
As per the literature survey works related to RESs Helal [19] proposed an EMS for a hybrid AC/DC MG for an
forecasting, most of the models are developed for short-term isolated community that employs a photovoltaic system for
and medium-term forecasting. It has been inferred that the desalination. The EMS of the MG is optimized with MILP
existing forecasting models used either statistical or machine resulting in daily operating costs minimization.
learning techniques to forecast RESs and some authors tried
hybrid models such as combining optimization / Umeozor and Trifkovic [20] proposed an EMS optimized
with MILP for a MG. Optimization is achieved with
decomposition methods with the algorithms. The authors also
recommended hybrid models are giving better performance parametrization of uncertainty factors of RESs and solved the
problems considering market price variations and detailed on
than individual models and the models are developed and
validated only for other countries not based Indian site storage system dispositions.
specific context. Xing [21] proposed a multi time-scale based EMS working
To cope with the complexity of the process and to improve on both static and dynamic programming models. Mixed
integer quadratic programming methods was employed for
the performance of the models, a novel hybrid ARIMA-
GARCH-NN approach is proposed based on the literature load flows and states of batteries were predicted with the
RESs availability.
review for the accurate RE prediction. The proposed hybrid
model combines both statistical and artificial intelligence Luna, [22] presented a real time EMS at different state of
techniques to forecast RESs. The historical data will be used operations such as perfect, imperfect and exact operations.
to train / develop the model with the site-specific measured The developed model has been tested with large imbalance
wind speed data, Solar Radiation and it also uses site-specific conditions between source and load both in Isolated and grid
wind/ Solar characteristics. connected mode.
V. ENERGY MANAGEMENT CONTROL SYSTEM Chaouachi [23] proposed an intelligent multi-objective cost
efficient EMS minimizing negative impacts on environment.
Management of the connection and disconnection events of
ANN is used for RESs prediction, EMS is optimized with
batteries and to shed low priority loads every time the load
multi-objective linear programming and battery scheduling
demand is greater than the generation capacity. This control
established with fuzzy logic protocols.
action anticipates any harmful operation of the system
through the predicting model, which predicts potential load Li., [24] proposed optimized EMS with particle swarm
imbalances; and to keep the voltage magnitude with a algorithm operating in both ON/OFF Grid modes. RE
maximum variation of ±5%. prediction approaches were employed for RESs and demand
side management to overcome the power quality issues.
Specific objectives are:
Marzband, [25] proposed EMS with artificial bee colony
(ABC) algorithm. A probabilistic approach is utilized for
 Effective functioning and co-ordination of developed operating the Gensets and intermittent RESs exaction
Microgrid system primarily aimed at achieving optimal resulting in 30% decrease in cost. Generation and load
resource scheduling. uncertainty operations were optimized with NN and markov
chain algorithms.
 Real time control and monitoring for predictive Farzin [26] proposed an EMS for isolated MG and used
maintenance stochastic approaches for minimizing the failures during
islanding process. The main objective of the papers is
 Uninterrupted energy generation from different RE efficient operational cost reduction of generation and the MG
sources and constant supply to local demand. system
Zhai[27] proposed a MG with Wind turbines, PV, batteries
 Minimize the operating costs of the microgrid and loads. The author presented a predictive robust control
algorithm for the standalone EMS and optimized with mixed
integer programming.
 Maximize the lifetime of energy storage systems
Solanki, [28] proposed a standalone MG and presented a
mathematical model for loads and EMS. Loads were
 Minimize the environmental costs. modelled with NNs and EMS is optimized with predictive
control methods resulting in optimal power dispatch.
Literature Review on EMCS Neves, [29] presented a comparative approach on EMS
optimization for standalone MGs. Mostly detailed on linear
Ahmad [17] proposed a techno-economic feasibility method programming and genetic algorithms proving that optimizing
to optimize the EMS of the MG with mixed integer linear the controllability on load could results in operational cost
programming (MILP). System sizing is done with Homer reduction.
software; Cost function is framed with linear programming
Inference International Symposium on Stochastic Models in Reliability
Engineering, Life Science, and Operations Management, SMRLO
In the presented literature review, different researches 2016, pp. 115–123. doi: 10.1109/SMRLO.2016.29.
have proposed several methodologies related to energy [7] Azad, H. B., Mekhilef, S. and Ganapathy, V. G. (2014) ‘Long-Term
management for an Isolated Microgrid. The literature review Wind Speed Forecasting and General’, Ieee Transaction on
details that multiple algorithms need to be employed in the Sustainable Energy, 5(2), pp. 546–553.
EMCS as the objectives and constraints are different. [8] Bali, V., Kumar, A. and Gangwar, S. (2019) ‘Deep learning based
wind speed forecasting- A review’, Proceedings of the 9th
Predominantly NNs is used for RESs prediction and Mixed International Conference On Cloud Computing, Data Science and
integer linear programming is used for optimizing the Engineering, Confluence 2019. IEEE, pp. 426–431. doi:
EMCS. Algorithms such as fuzzy logic works good for 10.1109/CONFLUENCE.2019.8776923.
battery sizing and state of charge (SoC) prediction, particle [9] Cadenas, E. and Rivera, W. (2010) ‘Wind speed forecasting in three
swarm algorithm for Genset and load operations, Genetic different regions of Mexico, using a hybrid ARIMA-ANN model’,
algorithm for operational cost reduction were also employed Renewable Energy. Elsevier Ltd, 35(12), pp. 2732–2738. doi:
for achieving stability of the Isolated MG operations. Hence 10.1016/j.renene.2010.04.022.
integrating an accurate RE power prediction algorithm with [10] Guo, Z. et al. (2012) ‘Multi-step forecasting for wind speed using a
modified EMD-based artificial neural network model’, Renewable
the EMS and implementing suitable optimization techniques
Energy. Elsevier Ltd, 37(1), pp. 241–249. doi:
for MG stability will result in sustainable energy solution for 10.1016/j.renene.2011.06.023.
the isolated community [11] Lei, M. et al. (2009) ‘A review on the forecasting of wind speed and
generated power’, Renewable and Sustainable Energy Reviews, 13(4),
VI. CONCLUSION pp. 915–920. doi: 10.1016/j.rser.2008.02.002.
[12] Mohandes, M. A. et al. (2004) ‘Support vector machines for wind
The diesel required for the generators needs to be transported speed prediction’, Renewable Energy, 29(6), pp. 939–947. doi:
10.1016/j.renene.2003.11.009.
from mainland and is dependent on the sea conditions.
During monsoons, the diesel transportation is extremely [13] Guo, Z., Dong, Y., Wang, J. and Lu, H., 2010. The forecasting
procedure for long-term wind speed in the Zhangye area.
challenging and cannot be transported in the extreme sea Mathematical Problems in Engineering, 2010.
states. The diesel generator sets are highly polluting and this [14] Barbounis, T.G., Theocharis, J.B., Alexiadis, M.C. and Dokopoulos,
is not a green form of energy. The existing diesel power plant P.S., 2006. Long-term wind speed and power forecasting using local
is located in a small populated area, thus compounding the recurrent neural network models. IEEE Transactions on Energy
effects of pollution on people. The growing needs of the Conversion, 21(1), pp.273-284.
power in Island can be augmented by development of [15] Nair, K. R., Vanitha, V. and Jisma, M. (2018) ‘Forecasting of wind
alternate energy systems. While the costs of fossil fuels are speed using ANN, ARIMA and Hybrid models’, 2017 International
on the rise, the cost of harnessing solar and other renewable Conference on Intelligent Computing, Instrumentation and Control
Technologies, ICICICT 2017, 2018-January, pp. 170–175. doi:
energy resources is consistently decreasing with technology 10.1109/ICICICT1.2017.8342555.
advancement. The dependence on RESs and its scalability [16] Shukur, O. B. and Lee, M. H. (2015) ‘Daily wind speed forecasting
make them well-suited to meet the power needs of remote through hybrid KF-ANN model based on ARIMA’, Renewable
areas that lack grid and road infrastructures. Energy. Elsevier Ltd, 76, pp. 637–647. doi:
10.1016/j.renene.2014.11.084.
Thus the Proposed system minimizes the dependence on
[17] Ahmad, J.; Imran, M.; Khalid, A.; Iqbal,W.; Ashraf, S.R.; Adnan, M.;
conventional sources and supports the energy needs with Ali, S.F.; Khokhar, K.S. Techno economic analysis of a wind-
RESs. By integrating the RE power forecasting algorithms photovoltaic-biomass hybrid renewable energy system for rural
with the EMCS along with real time control management electrification: A case study of Kallar Kahar. Energy 2018.
will paves for uninterrupted power supply making the [18] Taha, M.S.; Mohamed, Y.A.R.I. Robust MPC-based energy
intermittent resources as reliable and consistent. Thus management system of a hybrid energy source for remote
minimizing the total costs of generation, energy storage communities. In Proceedings of the 2016 IEEE Electrical Power and
Energy Conference (EPEC), Ottawa, ON, Canada, 12–14 October
satisfying the supply–demand constraints, ensuring 2016.
predictive maintenance and also protects the environment.
[19] Helal, S.A.; Najee, R.J.; Hanna, M.O.; Shaaban, M.F.; Osman, A.H.;
Hassan, M.S. An energy management system for hybrid microgrids in
REFERENCES remote communities. Can. Conf. Electr. Comput. Eng. 2017.
[20] Umeozor, E.C.; Trifkovic, M. Energy management of a microgrid via
[1] Wei Huanga, Miao Lua*, Li Zhangb. Survey on Microgrid Control parametric programming. IFAC-PapersOnLine 2016.
Strategies. In the proceedings of ICSGCE 2011: 27–30 September [21] Xing, X.; Meng, H.; Xie, L.; Li, P.; Toledo, S.; Zhang, Y.; Guerrero,
2011, Chengdu, China J.M. Multi-time-scales energy management for grid-on multi-layer
[2] Kutaiba Sabah , Hung Nguyen , Shantha Gamini , Jayasinghe , Thair microgrids cluster. In Proceedings of the 2017 IEEE Southern Power
Mahmoud A hybrid energy management and battery size optimization Electronics Conference (SPEC), Puerto Varas, Chile, 9 April 2018.
for standalone microgrids: A case study for Flinders Island, Australia , [22] Luna, A.C.; Meng, L.; Diaz, N.L.; Graells, M.; Vasquez, J.C.;
Article in Energy Conversion and Management · September 2018 Guerrero, J.M. Online Energy Management Systems for Microgrids:
[3] Ms. Jyoti B. Fulzele, Dr.M.B.Daigavane (PMME 2016). “Design and Experimental Validation and Assessment Framework. IEEE Trans.
Optimization of Hybrid PV-Wind Renewable Energy System” Power Electron. 2018.
Materials Today: Proceedings 5 (2018) 810-818-- Elsevier Science [23] Chaouachi, A.; Kamel, R.M.; Andoulsi, R.; Nagasaka, K.
Direct. Multiobjective intelligent energy management for a microgrid. IEEE
[4] Evans, Annette; Strezov, Vladimir; Evans, Tim (June Trans. Ind. Electron. 2013, 60, 1688–1699.
2009). "Assessment of sustainability indicators for renewable energy [24] Li, H.; Eseye, A.T.; Zhang, J.; Zheng, D. Optimal energy management
technologies". Renewable and Sustainable Energy -Elsevier Science for industrial microgrids with high-penetration renewables. Prot.
Direct. Control Mod. Power Syst. 2017.
[5] M. Liserre, T. Sauter, and J. Y. Hung, "Future energy systems: [25] Marzband, M.; Azarinejadian, F.; Savaghebi, M.; Guerrero, J.M. An
Integrating renewable energy sources into the smart power grid optimal energy management system for islanded microgrids based on
through industrial electronics," IEEE industrial electronics magazine, multiperiod artificial bee colony combined with markov chain. IEEE
vol. 4, pp. 18-37, 2010. Syst. J.2017.
[6] Ambach, D. and Vetter, P. (2016) ‘Wind speed and power forecasting
- A review and incorporating asymmetric loss’, Proceedings - 2nd
[26] Farzin, H.; Fotuhi-Firuzabad, M.; Moeini-Aghtaie, M. Stochastic [28] Solanki, B.V.; Raghurajan, A.; Bhattacharya, K.; Canizares, C.A.
Energy Management of Microgrids during Unscheduled Islanding Including Smart Loads for Optimal Demand Response in Integrated
Period. IEEE Trans. Ind. Inform. 2017. Energy Management Systems for Isolated Microgrids. IEEE Trans.
[27] Zhai, M.; Liu, Y.; Zhang, T.; Zhang, Y. Robust model predictive Smart Grid 2017.
control for energy management of isolated microgrids. IEEE Int. [29] Neves, D.; Pina, A.; Silva, C.A. Comparison of different demand
Conf. Ind. Eng. Eng. Manag. 2018. response optimization goals on an isolated microgrid. Sustain. Energy
Technol. Assess. 2018.