Energy Storage System Scheduling in Wind-Diesel Microgrids Michael Ross Department of Electrical & Computer Engineering McGill University Montr´ eal, Canada July 2010 A thesis submitted to McGill University in partial fulfillment of the requirements for the degree of Master of Engineering. c 2010 Michael Ross
95
Embed
Energy Storage System Scheduling in Wind-Diesel Microgridsdigitool.library.mcgill.ca/thesisfile95237.pdf · Energy Storage System Scheduling in Wind-Diesel Microgrids ... July 2010
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Energy Storage System Scheduling in
Wind-Diesel Microgrids
Michael Ross
Department of Electrical & Computer EngineeringMcGill UniversityMontreal, Canada
July 2010
A thesis submitted to McGill University in partial fulfillment of the requirements for thedegree of Master of Engineering.
c� 2010 Michael Ross
i
Abstract
This thesis proposes a knowledge based expert system tool that can be used as an on-
line controller for the charging/discharging of an energy storage system in a wind-diesel
microgrid. The wind-diesel microgrid is modelled, and a typical energy storage system
is implemented to test the functionality of the controller using hourly-discrete power val-
ues. The results are compared against an offline optimization that was provided 24-hour
lookahead wind values, as well as a controller that was implemented using artificial neural
networks. The knowledge based expert system is then used to analyze the cost of energy,
by means of a parametric analysis, consisting of varying the wind penetration, energy stor-
age system power rating and energy rating to determine for which wind penetration values
a storage system implementation would be technically and economically viable. Differ-
ent storage technologies are tested in a one-year time frame to determine which would be
best suited for this particular application. The energy storage systems are implemented
as single-layer and dual-layer, in which the knowledge based expert system is modified for
the latter analysis, in order to determine whether or not there are advantages to having
a dual-layer storage system. Throughout these analyses, the flexibility of the knowledge
based expert system controller to various energy storage systems and microgrid models
is verified. It also demonstrates that, in a context of high base generation costs, energy
storage can be a viable solution to managing wind power variations.
ii
Resume
Cette these propose un systeme expert avec une base de connaissance qui peut etre utilise
comme un contoleur lors de la charge et de la decharge d’un systeme de stockage d’energie
dans un micro-reseau eolien-diesel. Un micro-reseau eolien-diesel modele est etabli, et un
stockage est installe pour tester les fonctionnalites du controleur en utilisant des valeurs
de la puissance horaire. Les resultats sont compares avec une optimisation utilisant 24
heures de valeurs en avance pour la vitesse du vent, et aussi avec un controleur base
sur un reseau de neurones artificiels. Le controleur systeme expert est ensuite utilise pour
analyser les couts d’energie d’une analyse parametrique, en variant la penetration du vent, la
puissance nominale du stockage, et la capacite nominale du stockage. Cette analyse indique
pour quelles valeurs de penetration eolienne une mise en œvre d’un stockage serait viable
economiquement et techniquement. Differentes technologies de stockage sont testees afin de
determiner laquelle serait le mieux adapte pour cette application particuliere. Les systemes
de stockage sont realises a l’aide d’un ou de plusieurs types de systemes, et le controleur
systeme expert est modifie en consequence, afin de determiner s’il y a des avantages a avoir
ce type de stockage. Ces analyses montrent aussi que le controleur systeme expert a la
capacite et la flexibilite de s’adapter a des technologies ainsi qu’a des micro-reseaux de
differents types.
iii
Acknowledgments
I would like to sincerely thank my supervisor, Professor Geza Joos, for his guidance and
support throughout my Master’s studies. I have learned so much in my education at McGill
University, and I owe a large part of that to Professor Joos. He challenged me with high
objectives and pushed me to always strive to be better, and he was always able to reward
the hard work. I am extremely grateful for the opportunities that he has provided me and
allowed me to pursue. I am excited to continue my studies under his supervision.
I would also like to thank my colleagues in the Electric Energy Systems Laboratory
group. In particular, I would like to sincerely thank Rodrigo Hidalgo for his help through-
out my Master’s program. He has co-authored most of my publications and has helped
in teaching me many things about practical issues of power systems. He has been a close
friend that I can always trust and rely on. My gratitude also goes out to Dr. Chad Abbey
for his help and guidance throughout my research endeavours, by taking the time to help
and explain many things to me, and for making his LATEX thesis template available to
me. Thanks to Mohamed El Chehaly, Carlos Martinez, Jonathan Robinson, Amir Kalan-
tari, Etienne Veilleux, and Hamed Golestani Far, and other people in the Electric Energy
Systems Laboratory for being helpful and giving advice on how to improve my research.
I would like to thank my family and friends that have helped me keep a balanced life.
I know that my family is there for me whenever I need it, and just that knowledge has
enabled me to take chances in life and strive to be a better person. My parents are the
most selfless people I know, and their constant support has enabled us to grow into a strong
and close family. I can always count on my siblings, Stephanie, Nicole, Katie, Bryan, and
John, to be there for me if ever I need to relax, have fun, or just joke around. I would
particularly like to thank Bryan, who was my housemate throughout my time in Montreal.
It’s always a great pleasure spending time with my family.
Finally, I would like to thank the friends I made through the McGill Rowing Team,
McGill Nordic Skiing Team, and the McGill Triathlon Team. Because of everyone men-
tioned here, I feel I have been able to keep my mind, body and soul in proper balance.
AI Artificial IntelligenceANN Artificial Neural NetworkCAES Compressed Air Energy StorageDER Distributed Energy ResourceES Expert SystemESS Energy Storage SystemGAMS General Algebraic Modelling SystemIEEE-RTS Reliability Test System-1996KB Knowledge BaseKBES Knowledge Based Expert SystemMLC Minimum Loading ConstraintO&M Operation and MaintenanceRE Renewable EnergySMES Superconducting Magnetic Energy StorageSOC State of ChargeVRB Vanadium Redox Flow BatteryWTG Wind Turbine Generator
xi
List of Symbols
The following lists the most important symbols used in the thesis. Where relevant, boldface
is used to denote vectors or matrices.
Indices
h Index of total time periods
t Index of time periods from 1 to T
Functions
Cy(·) Yearly cost of operating the system [$]
ckWh(·) Average minimized cost of energy per kilowatt-hour [$/kWh]
Parameters
PESS Power rating of ESS [kW]
PESS,m Power rating of medium-term ESS [kW]
PESS,s Power rating of short-term ESS [kW]
Pmin Diesel minimum loading constraint [kW]
Pd,max Maximum diesel generating power [kW]
PW Power rating of wind farm [kW]
PL Base load power rating [kW]
xii List of Symbols
EESS Energy rating of ESS [kWh]
EESS,m Energy rating of medium-term ESS [kWh]
EESS,s Energy rating of short-term ESS [kWh]
Et−1 Energy state of ESS at the beginning of hour, t-1 [kWh]
NWTG Number of wind turbine generators
πd Cost of energy supplied by the diesel generator [$/kWh]
πw Cost of energy supplied by the wind farm [$/kWh]
πOMfFixed O&M Costs of ESS [$/kW/year]
πOMv Variable O&M Costs of ESS [$/kW/year]
πPCS ESS Power Conversion System initial cost [$/kW]
πdis ESS disposal cost [$/kW]
πess,p Incremental cost of ESS storage power rating [$/kW]
πess,e Incremental cost of ESS storage energy rating [$/kWh]
πESS Total annual average cost of ESS [$/year]
ηch Charging efficiency of ESS [%]
ηdis Discharging efficiency of ESS [%]
η Round-trip efficiency of ESS [%]
rwl Wind power penetration [%]
L Average lifespan of ESS [years]
T Time period that is being analyzed [h]
Variables
PL,t Community load power demand for hour t [kW]
Pw,t Wind power production for hour t [kW]
Pres,t Residual power of load minus wind for hour t [kW]
Pres,max Maximum excess power that the ESS would have to absorb or deliver
[kW]
pess,t Power of ESS during hour t [kW]
pess,m,t Power of medium-term ESS during hour t [kW]
pess,s,t Power of short-term ESS during hour t [kW]
pdiesel,t Diesel generator power during hour t [kW]
List of Symbols xiii
pdump,t Dump load power during hour t [kW]
Eloss Total lost energy due to inefficiencies of charging and discharging the
ESS [pu]
Edump Total lost energy wasted through the dump load [pu]
Eres,max Maximum excess energy that the ESS would have to store [kWh]
ESStech ESS technology used
eess,t Energy in ESS after hour t [kWh]
eess,m,t Energy in medium-term ESS after hour t [kWh]
eess,s,t Energy in short-term ESS after hour t [kWh]
udiesel,t Binary variable associated with diesel plant dispatch at hour t (0 = off,
1 = on)
x Vector of free variables in the minimization function
xiv
1
Chapter 1
Introduction
1.1 Background
Remote communities are typically isolated from large power systems (off-grid), and there-
fore must supply their load through their own means. Some of these remote communities
depend mainly on fossil fuel power plants for their energy supply. Many microgrids use
diesel generation to feed the local grid, however this results in inherently high operation
costs and emissions. These negative effects can be reduced with the integration of Renew-
able Energy (RE) sources at a high penetration level [1].
In recent years, in Canada as well as in other countries, the use of RE has grown as a
source of electricity in many power systems. This is largely due to government incentives
and public pressures to find alternatives to carbon-emitting generators. The aforementioned
remote communities may benefit greatly from the integration of RE generation to their
micro systems. Such integration would reduce the costs of fossil fuel consumption while
enhancing the sustainable development of the area, thus providing both economical and
environmental advantages [2]. For instance, some remote communities are located in an
area of high winds, thus making it economically justifiable to incorporate wind generation
into their microgrid. However, power systems are not designed to operate completely on
RE generation due to the inconsistent and intermittent source of power, and so it is still
dependent on fossil-fuel based power generation [3].
Planning and operating power systems with medium to high RE penetration levels im-
plies some difficulties due to the stochastic and intermittent characteristics of the sources
[4–6]. Exact power production of wind generators is irregular and difficult to accurately
2 Introduction
predict. Furthermore, it is necessary for isolated grids with fossil-fuel based power genera-
tion that have been integrated with RE (primarily wind farms) to implement dump loads
or generation curtailment [3]. This practice is in place to adequately meet the demand and
keep the power balance while operating within the technical constraints of the diesel plant,
as the diesel generator must operate above a Minimum Loading Constraint (MLC). The
use of dump loads or generation curtailment wastes this excess generation and reduces the
benefits of RE, which in turn affects the investment in such technologies.
The implementation of an energy storage system (ESS) may be beneficial, as it may
help to increase the RE penetration in the system [7]. An ESS is able to store the excess
generation during times of high wind/low load conditions, and deliver the power during
low wind/high load conditions. An ESS with a large storage capacity may offer a suitable
alternative to wasting energy from intermittent sources, and they increase the penetration
of RE [2, 5, 7]. In addition, an ESS that includes power conversion equipment and control
also offers operational advantages, such as power fluctuation suppression (power smoothing)
and voltage and frequency regulation [8,9]. Energy that would have otherwise been wasted
can be used later to provide power instead of increasing the diesel output. In this sense,
the total diesel consumption, carbon emissions, and fuel costs will decrease, thus increasing
the value of the RE investment and improving the output controllability [7, 10]. Another
advantage to implementing an ESS is that the system can operate in discontinuous diesel
mode, which may reduce the amount of diesel that is consumed.
Despite these advantages, there are also some disadvantages to ESSs. Primarily, they
are associated with a very high capital cost, which may not make them a financially vi-
able solution to the microgrid. These high implementation costs associated with the ESS
makes it necessary to perform an economic analysis in conjunction with any initial techni-
cal analysis considering the implementation of an ESS to determine the feasibility of the
project. An appropriate estimate on the return on investment is based on the advantages
that an ESS could provide to the community in terms of reducing fuel consumption and
optimization of the use of the wind power [2]. Another issue involved with ESSs is its
technical feasibility. Some ESSs fare better with longer-term storage, but do not handle
fast charging well, while others can do fast charging but may not be able to provide a high
energy capacity [11]. Further analysis is also required for initial estimates of the sizing and
costs of ESS implementation [12,13].
Scheduling of the available resources is required for optimal grid performance to meet
1.2 Literature Review 3
the load and reduce costs and emissions. Several studies have been conducted to optimize
the operation and scheduling of power systems with an integrated ESS [4–7]. This thesis
presents a controller that is implemented as a Knowledge Based Expert System (KBES),
with an hourly-discrete scheduling algorithm developed for an isolated power system with
diesel-wind-ESS resources. The objective of this KBES is to optimize the cost of opera-
tion by minimizing the use of dump loads (wasted energy), and therefore reducing diesel
consumption, assuring that the demand and all other constraints are met. The KBES tool
is able to control the diesel generation and the charging/discharging cycles of the storage
system from the wind and load profiles one hour in advance.
1.2 Literature Review
1.2.1 Microgrids
Microgrids are self-contained power systems that are isolated from large power systems.
Therefore, they must provide energy to their local loads themselves through at least one
distributed energy resource (DER), but they cannot provide power to other loads on a
different grid. Typically, they are located too far from a large grid to enable them to be
connected.
Distributed generators can be of many different types. Synchronous generators and
induction generators are very typical, and they act as the interface between the power
source and the rest of the microgrid [14]. Power electronic interfaced DERs, however, can
also provide voltage and frequency control. The control of these units depends on whether
or not their power source can be dispatched; e.g., wind farms are typically controlled for
peak power output. If the microgrids cannot be connected to another grid, load/generation
shedding must be used to maintain the power balance. The control of the power system
can either be centralized by one actor, or by using a decentralized approach [15].
When using more than one generator of different ratings or technologies, the scheduling
of the power output can be controlled to minimize fuel consumption [16]. The load can be
shared by one or both generators (if they are rated appropriately to do so), and the cost of
energy can be minimized. A communication link between the two power sources will facili-
tate the optimization process. The selection of the different DERs, as well as implementing
an energy storage system, will have an effect on the operation of the microgrid, and some
4 Introduction
technologies can provide operational advantages such as power response time and system
stability improvements [17].
In [18], the management issues of integrating RE generation is assessed based on its in-
tegration in a university microgrid. This paper uses Information and Communication Tech-
nology to help manage the operation of the system. A small variable speed wind turbine is
modelled as its RE generator, and it was shown that the information and communication
technology aided the human-interfaced system operation to maintain the power balance
and stability throughout its operation. Another control method, implemented in [19], uses
real power droop to regulate the power output of the DER in a microgrid. The objec-
tive of the control is to adjust its generating output to minimize fuel consumption while
maintaining stable operation.
In some situations, the power system operation can be evaluated with a microgrid that
is able to connect to a larger power system. In this case, the DER can be used to supply
the load in the microgrid by itself or with the help from the grid, or provide additional
power to the grid; it also has the advantage of providing local power reliability [20]. While
connected to the power grid, the microgrid must control its system taking into consideration
economic scheduling, load forecasting, security of the system and demand side management
functions.
1.2.2 Energy Storage Systems
There are many advantages to incorporating an energy storage system into a microgrid.
For instance, ESSs can be used to optimize the energy from RE sources and reduce fuel
consumption, offering economic and environmental advantages. When implementing wind
generation, the random nature of this DER can lead to issues with both probabilistic and
deterministic criteria of the planning and design of microgrids [21]. An ESS can help
alleviate these issues, by providing control to such uncontrollable resources.
Energy storage systems can ameliorate microgrid operations by improving the relia-
bility and quality of generated power, support other DER, and time shift generation of
uncontrollable sources [22]. When integrated with wind generation, it can also provide
the system reserve, increase the overall system operation efficiency, enhance wind power
absorption, provide fuel cost savings, and reduce CO2 emissions [23]. In [24], an ESS is
used to alleviate the intermittency of the wind generation in a microgrid and reduce the
1.2 Literature Review 5
Fig. 1.1 Typical ESS power (x-axis) and energy (y-axis) capabilities [26].
amount of wind curtailment. It was shown that an ESS can reduce the power loss and
improve the voltage profile in a system with a high wind power penetration, to the extent
that an ESS incentive program should be recommended to the regulators.
There are many different ESS technologies available [22, 25], and Fig. 1.1 shows the
typical ESS ratings of different technologies. Battery technologies have been shown to pro-
vide frequency control and stability for longer duration requirements [25, 27]. In [28], a
superconducting magnetic energy storage system is used to perform both power flow and
damping enhancements of a large wind farm. It was found that the system successfully sta-
bilizes the power output and reduces the power fluctuations inherent with wind generation.
In [29], Direct Methanol Fuel Cells are modelled based on statistical design of experiment
methodology to detail their complex nonlinear functionality. Other examples of particular
ESS technology uses include, but are not limited to, flywheel energy storage and Vanadium
Redox Flow Batteries [30, 31].
Many papers address the economic issues involved with an ESS. A nonlinear optimal
6 Introduction
ESS scheduling algorithm is presented in [32] that optimizes the charging and discharging of
the ESS using time-of-use rates with wind energy in the system. This controller determines
the optimal operating values for the next time step, and it can be used in real time and
for offline analyses. In [33], an analysis based on a Dutch generation system with large
wind generation is simulated for an implemented ESS. The operational cost savings from
the ESS increases with an increase in installed wind capacity, however it was deemed that
the large initial investment may not be worth the cost savings. In [34], attractive value
propositions for modular electricity storage are outlined. Among these propositions include
the improvement of local power quality, and electric utility end-user cost management
during times of peak or critical peak pricing.
1.2.3 Expert Systems
Many papers describe methods of using Expert Systems for control purposes. Some meth-
ods include Data Mining, Genetic Algorithms, Artificial Neural Networks, Fuzzy Logic,
and Knowledge Based Expert Systems.
Data mining is used to discover knowledge within a large amount of data [35]. In [36],
data mining is used to predict the power output of WTGs based on wind speeds that are
evaluated at different time intervals. It was found that although the wind power curves
may not be completely reliable in determining the total wind farm power output, the data
mining technique predicts the wind power very accurately when provided a multi-period
lookahead.
Genetic algorithms use populations of results, and by combining them and keeping the
best solutions for that generation, a solution may be achieved after several iterations. This
has been used for scheduling and planning in short-term and long-term time intervals.
In [37], genetic algorithms are used to plan the future expansion of an electric distribution
system over the following seven years. The algorithm uses a dynamic programming model,
whose design actions are interdependent and complex, that was tested on a 100-bus dis-
tribution system. Although it yielded higher initial costs than other expansion scheduling
approaches, the final costs were much lower after several generation iterations. In [38],
a genetic algorithm with multiplier updating is implemented to determine the optimized
power for an economic load dispatch of complex systems. The benefit of using genetic
algorithms is that mathematical models may not be applicable to solve this problem. This
1.2 Literature Review 7
has been found to obtain a minimum cost and requiring only a small amount of populations
for real-world economic dispatch problems.
Artificial Neural Networks (ANN) employ adaptable nodes that can modify their firing
threshold to output desired results. ANNs have been successfully applied to power systems
in [39], which used them for load forecasting, fault classification, voltage stability, economic
dispatch, and power system stabilizer design. In [40], a fast-learning ANN is used as a real-
time fault detector for distance protection. Another use in power systems is to predict
different transformer oil parameters, which was implemented using a three-stage neural
network cascade in [41].
Fuzzy logic implements certainty weightings associated with logic values, such that
they can have a value between 0 and 1 instead of being strictly binary. Fuzzy logic is
used to control the coordination of a circuit breaker with an implemented superconducting
magnetic energy storage system (SMES) in [42]. It was found that the fuzzy logic controller
performed much better in terms of system stability than by simply auto-reclosing the circuit
breaker, and the control system also had a better transient stability performance than a
static nonlinear controlled SMES. In [43], a fuzzy logic controller is used to control the power
output and voltage stabilization of a four-machine power system. This paper also uses a
systematic analytical method so that there is no need of prior knowledge of the system.
Results show that the controller is able to provide good system stability and sufficient
oscillatory damping without compromising voltage regulation. In [44], fuzzy optimization
models are used to determine day-ahead unit commitment of generators in power systems
with wind generators. Fuzzy set theory helps to find an optimized solution to imprecise or
conflicting objectives.
Conversely, this thesis focuses primarily on a Knowledge Based Expert System (KBES).
The purpose of developing a KBES is to create an engine that attempts to perform tasks
at the level of experts in the respective domain (i.e. a system that imitates the problem-
solving behaviour used by human experts [45, 46]). The general structure of an expert
system includes the user inputs, the Knowledge Base (KB), an inference engine, and the
outputs. The inference engine combines the user inputs with the knowledge base to generate
the outputs [47].
The facts and the rules that constitute the expert system are derived from one or
several human experts in the problem domain [48]. In order for the KBES to perform as
an expert, it is necessary that both declarative and procedural knowledge possessed by
8 Introduction
experts are implemented in the KB. Declarative knowledge is concerned about the facts
and concepts, as well as the relationships between them within in a particular domain.
Procedural knowledge describes how to reason or handle the information within the KB
[45,46].
The interface of the expert system should be developed so that it could be used to solve
problems even by users without experience in the knowledge domain [49]. The Knowledge
Engineer (designer) must to attempt to develop a user-friendly system in order to avoid
wasting the user’s time and effort struggling with how to use it. In the design process, all
the user requirements should be considered for a successful system design [46].
1.3 Research Objectives
1.3.1 Problem Definition
The purpose of this thesis is to develop a controller that will optimally schedule the charg-
ing/discharging cycle of the ESS and the diesel generation to cover the power balance
within the system’s constraints. This controller will be used to determine under which
circumstances an ESS is technically and economically justifiable, and which technologies
are best suited for this application.
1.3.2 Contributions
This thesis provides new research in the area of scheduling and operation of a wind-diesel-
ESS microgrid. In particular, it provides insights into the implementation phase of incor-
porating an ESS through the use of a Knowledge Based Expert System Controller. Several
contributions of this thesis include:
• The creation of an online controller, implemented as a Knowledge Based Expert Sys-
tem, that is able to dispatch the diesel generation and the ESS charging/discharging
that can yield optimized results for minimizing the dump load consumption and the
cost of energy. These results have been presented and published in [13].
• A parametric analysis of the wind penetration and ESS ratings in the microgrid to
determine for which penetration values and for which ESS sizes would yield the lowest
1.4 Summary 9
cost of energy and best technical operation of the system. These results have been
presented and published in [50].
• An analysis of microgrid operation under both continuous and discontinuous diesel
modes of operation. These results have been presented and published in [51].
• An analysis of various ESS technologies in single-layer and dual-layer configurations.
These results have been presented and published in [52].
Work performed in Chapters 2 and 3 have been done in collaboration with fellow grad-
uate student Rodrigo Hidalgo Anfossi. He has provided much of the research data into the
system parameters and he helped develop the KBES Controller. He has also been consulted
on further work to verify the logic of the methodology and knowledge base.
1.4 Summary
This thesis is divided into the following chapters:
Chapter 1: Introduction
In this chapter, an introduction to the problem area is presented. Some generic, key
problems are addressed, and a literature review section details what research has been
done in this area in an attempt to solve the problem as well as overviewing other AI
techniques for control. The purpose of the thesis, in terms of the research goals and the
contributions, are given in the Research Objectives.
Chapter 2: System Model and Knowledge Based Expert System Controller
This chapter will detail the base model within which the Knowledge Based Expert System
tool will control. The basic parameters, the technical constraints, and the mathematical
models of each component are outlined. Once this is established, a description of the KBES
Controller is given with the rule base that is derived from the model system.
Chapter 3: Validation of the KBES Controller
In this chapter, the KBES Controller is tested online, and the results of the testing are
provided. The success of the controller is based upon the amount of diesel fuel that is
10 Introduction
saved, measured by the reduction in diesel generation, and the reduced cost of energy to
the community. These results are compared against an offline optimization function, as
well as to an Artificial Neural Network controller that is attempting to achieve the same
goal.
Chapter 4: Parametric Analysis of Varying Wind Penetration and ESS Sizing
This chapter modifies the microgrid model developed in Chapter 2 to analyze the effects of
varying the level of relative wind generation and ESS power rating and energy rating. The
purpose of this chapter is twofold. First, it demonstrates the flexibility and adaptability of
the KBES tool. Second, it shows the effect of implementing various ESS sizes in microgrids
of different wind penetrations on the cost of energy. This chapter also determines at which
penetration level the implementation of an ESS is financially justified for both continuous
and discontinuous diesel operation.
Chapter 5: ESS Technology Analysis Using KBES Tool
In this chapter, different ESS sizes of different ratings are implemented in the base micro-
grid model from Chapter 2. At first, a single-layer ESS is implemented for ten different
technologies for continuous and discontinuous diesel operation. Then, the KBES is mod-
ified to enable a dual-layer ESS, where medium-term ESS technologies are coupled with
short-term ESS technologies. The technical and economic analyses for these results are
provided.
Chapter 6: Conclusions
This chapter concludes the thesis by reviewing the work that has been done and explicitly
details the conclusions that are drawn from these analyses. This chapter is divided into
subsections to summarize these results for Chapters 2-5. Future work is then suggested in
order to guide the advancement of research in this area.
11
Chapter 2
System Model and Knowledge Based
Expert System Controller
2.1 Introduction
This chapter will cover the system model for which the Knowledge Based Expert System
Tool will control the energy generation, dump, and charge/discharge of the ESS. This model
will be used as a base model, and studies detailed in further chapters will use this model
or a modification. The KBES Controller is also detailed, and is based on the parameters
of the microgrid.
2.2 Microgrid Model
The isolated power system, shown in Fig. 2.1, comprises a diesel generator, the community
load, a wind farm, an ESS, and a dump load. Transmission losses and reactive power flow
are neglected since a small remote system is considered, although one method of how it
can be implemented is presented in [53]. Community load and wind data are assumed to
be deterministic for the following hour; that is, the forecasting of these values are assumed
to be known without errors. The impact of the energy storage sizing, and the wind and
load forecasting accuracy, which is discussed in [6], is out of the scope of this thesis.
The main goal of the KBES Controller is to minimize the cost of the operation of the
system over a given period. The wind and load data are independent and are treated
as uncontrolled inputs. It is assumed that the ESS has been previously sized using [13].
12 System Model and Knowledge Based Expert System Controller
Fig. 2.1 Wind-diesel power system with an energy storage system.
The controller must then schedule the diesel generation and the charging and discharging
of the ESS in order to minimize the cost of operation, hour by hour, based on the rules
implemented in its Knowledge Base (KB).
The yearly cost of operating the system can be calculated as:
Cy(x) = πess,e · EESS + πess,p · PESS +T�
t=1
(πd · pdiesel,t + πw · Pw,t) (2.1)
which is then divided by the total annual load to yield the cost per kilowatt-hour.
The costs associated with the ESS implementation and Operation & Maintenance
(O&M), πess,e and πess,p, include the incremental costs, fixed costs and variable costs of
O&M of the ESS throughout the projected lifetime.
2.2 Microgrid Model 13
The optimization problem is defined as:
minx
Cy(x) (2.2)
where the vector of free variables x is:
x =�
pdiesel pdump pess udiesel
�(2.3)
subject to the constraints of the power system characteristics as described in Sections 2.2.1
through 2.2.4 and the power balance.
The optimal results are found through an optimization function implemented in GAMS
(General Algebraic Modelling System) for (2.2), with the constraints listed in Sections
2.2.1-2.2.4. Note that this is minimizing the generation from the diesel generator since the
costs associated with the ESS in (2.1) are fixed for this analysis, and the wind generation
is also pre-defined.
The GAMS optimization is used as a benchmark against which the KBES Controller
is compared. The twenty-four future hourly values for wind and load are provided to
the GAMS optimization in order to appropriately schedule the power for the given hour.
Conversely, the KBES scheduling implementation is only given the values for the current
hour. The KBES is implemented in this manner to justify whether it is necessary to provide
forecasting values for ESS scheduling systems, as used in [6]. Since it is not provided with
future data, the KBES attempts to minimize the diesel generation by minimizing the power
wasted through the dump load for every hour. Formally written, the KBES attempts to
minimize:
minx
pdump,t ∀t ∈ T (2.4)
with the same constraints as (2.2). Section 2.3 details the implementation of the KBES.
14 System Model and Knowledge Based Expert System Controller
2.2.1 Diesel Generation
The technical characteristics of the diesel generator used in this analysis are extracted
from [54], and is given in Table 3.2. One important factor in the operation of diesel-fuelled
equipment is their Minimum Loading Constraint (MLC). This should be considered in order
to avoid premature aging and engine failure [3]. In this case, a 30% MLC of the generator
rating was defined:
Pmin = 0.3 · Pd,max (2.5)
Two operating cases are considered in this thesis: one with the diesel generator operating
continuously, and the other that allows for the diesel to be turned off. For both cases, a
binary variable will be defined as:
udiesel,t =
�1 diesel generator turned on
0 diesel generator turned off(2.6)
where udiesel,t = 1 ∀t ∈ T for the continuous case. Although there are both time and
financial costs associated with inefficiencies of turning on and off the diesel generator, these
are not factored into this optimization problem. Ramping constraints of diesel generators
are also neglected, so the generator can instantly increase or decrease its power output to
any desired level within its operation constraints, which are defined as:
fide Batteries, and Nickel-Cadmium Batteries. Note that the Sodium Polysulfide Batteries
performed much worse for the discontinuous diesel operation, due to the high initial power
rating cost. This demonstrates that the best technologies to use does depend somewhat on
the diesel operating mode. In all cases, the Lead Acid Batteries performed the best and
the Nickel-Cadmium Batteries performed the worst in both the technical and economic
analyses.
This chapter also shows that the KBES can be used to perform parametric analyses
of various ESS technologies and configurations. It can therefore also be used as a sizing
method to find the best ESS technology and sizes for including it in a wind-diesel microgrid.
67
Chapter 6
Conclusions
6.1 Thesis Summary
A knowledge based expert system controller is used to schedule the diesel generation and the
energy storage system charging and discharging in a wind-diesel-ESS microgrid to minimize
the cost of energy. The KBES Controller comprises a set of rules within its knowledge
base, which are constructed from the system constraints. These rules can be fired at any
time, and once all the constraints are met (including maintaining the power balance in
the system) the dispatch for the next hour is determined. The KBES Controller aims to
minimize the use of the dump load that is normally associated with diesel operation. The
results are compared to an offline optimization algorithm applied to the same power system
and ESS size that has a 24-hour lookahead. The results obtained show that by minimizing
the energy wasted through the dump load with the use of the ESS and KBES Controller,
the required diesel generation is reduced, therefore reducing operation costs and emissions.
Another benefit of the KBES Controller s that only near-term wind forecasting is required,
not medium- or long-term. This is beneficial since long-term wind data forecasting may
not be available or accurate.
This KBES tool has been shown to be very flexible, as it was used to perform various
analyses of wind-diesel-ESS microgrids. At first, a parametric analysis is performed where
the wind penetration and ESS ratings were modified to determine its effect on the system
operation and cost of energy. The purpose of this study was to determine for which wind
penetrations an ESS implementation would be most beneficial to the microgrid. Both
continuous and discontinuous diesel operations are analyzed with the result that, in general,
68 Conclusions
an ESS would yield a larger return on investment (and hence, a lower cost of energy) if
discontinuous diesel operation were used.
Different ESS technologies and configurations are also analyzed as there are many avail-
able that can offer financial or technological advantages. In this analysis, the system is sim-
ulated for one year for each technology, with the implementation costs and lifetime taken
into consideration. The overall results are analyzed for the single-layer ESS and the dual-
layer ESS, with fast-charging, costly ESSs used as the short-term ESS, and other (mostly
battery) technologies used as the medium-term ESS. Some ESS technologies have been
shown to be a technically and economically viable solution to wind-diesel microgrids. Ad-
ditional benefits of using an ESS is that it may increase the lifecycle of the diesel generator
since it is used less frequently at a higher operating point.
These analyses verify that the KBES Controller can be used for different microgrid
systems and for different ESS technologies, as long as the rules are implemented such that
they account for the constraints of the system. By using this tool in various test systems,
it was found that an implementation of an ESS can not only help with balancing the power
and increasing the amount of RE penetration in the system, but can furthermore decrease
the overall cost of energy.
6.2 Conclusions
The results and conclusions of the various chapters in this thesis are summarized below.
Chapter 2
In this chapter, a wind-diesel microgrid is modelled. The mathematical models of the sys-
tem are defined and implemented so that a KBES Controller can be integrated online. The
KBES Controller controls the scheduling of the diesel generator, ESS charging/discharging,
and dump load consumption using one-hour lookahead values. It comprises rules within its
knowledge base that are derived simply from the system constraints to minimize the cost
of energy in the system.
6.2 Conclusions 69
Chapter 3
This chapter demonstrates that the proposed Knowledge Based Expert System Controller
yields optimized results when attempting to minimize fuel costs in the isolated wind-diesel-
ESS power system. By minimizing the power lost through the dump load, the diesel
generator power is reduced, and it thus minimizes the generation cost during the projected
lifetime for the chosen system. Both the continuous and discontinuous diesel generation
modes are considered, and the respective results are compared to those obtained by using an
offline optimization function which is allowed a 24-hour lookahead window for each hour.
Both modes of operation are shown to have similar results to the offline optimization,
obtaining results that differ by less than 0.3%. Therefore, it is sufficient to have only a
one-hour lookahead and obtain near-optimal results using a KBES Controller for energy
storage for the proposed system.
Chapter 4
In this chapter, the KBES Controller is used to perform a parametric analysis to determine
the effects of varying the wind penetration, ESS power rating and ESS energy rating. The
discontinuous diesel mode has a large effect on the optimized size of the ESS as well as the
manner in which the system dispatches its controllable generators and loads. A large ESS
size will allow for more energy to be stored, but it has a high initial cost. For continuous
diesel operation, there is a minimum cost difference between implementing an ESS and
not implementing an ESS around rwl ≈ 0.7-0.8. For the discontinuous case, however, the
cost difference continued to decrease with an increased wind penetration, implying that it
is more economical to implement an ESS in discontinuous mode as the wind penetration
increases. A sizing methodology can be derived from the parametric analysis to determine
which ESS ratings will yield the minimum cost of energy for a particular microgrid.
Chapter 5
A technical and economical analysis of varying the ESS technology is performed in this chap-
ter to determine which technology is best suited for the microgrid model. The methodology
is established to perform a parametric analysis of the different ESS sizes, and from the one-
year analyses, determine the lowest cost of energy. For the single-layer ESS analysis, it was
found that the Lead Acid Batteries performed the best, yielding a higher power and energy
70 Conclusions
rating and the minimum cost of energy when compared to the other technologies. Other
technologies that performed well were (in order from best to worst) Sodium-Polysulfide Bat-
teries, Sodium-Sulfur Batteries and Zinc-Bromine Batteries. For all other ESSs analyzed, it
was determined that in order to obtain a minimized cost of energy, the best solution would
be to not implement those technologies. For the dual-layer ESS analysis, it was determined
that it is not economically viable to implement a short-term ESS in this hourly discrete
analysis. When analyzing the cost of energy for the various power and energy ratings of
the Lead-Acid ESS, differences became clear between the continuous and discontinuous
diesel mode of operation. For the continuous case, a gradient can be observed in the cost
of energy when increasing both the power and energy ratings, with a higher sensitivity to
the power rating. For the discontinuous case, the cost peaks at a certain power rating, and
then decreases and settles for higher power ratings. This is due to the fact that at this
point, the diesel generator can take advantage of its discontinuous operation. There is also
relatively little sensitivity to the energy rating on the cost of energy for the discontinuous
diesel mode.
6.3 Future Work
The next step in the implementation of the KBES Controller would be to perform a Hard-
ware in the Loop (HIL) simulation [71]. This would justify its functionality in real time
with an emulated microgrid as opposed to simply using discrete, hourly values. In real-time
operation, special considerations must be given to stand-alone operation of wind-ESS with
regards to power and voltage control of the system, as presented in [72]. Or, if the wind
data is not available, even taking discrete time variables of smaller periods of time would
allow for the short-term wind power fluctuations to be modelled, which could have a drastic
effect on how the ESS is utilized.
In order to not rely on a user creating the rule base for the microgrid, data mining can be
used as a front end of creating the rules [73]. Data mining will allow the Controller to learn
the specific constraints of the system and schedule the generation and ESS appropriately
through the generated boundary conditions. One benefit of using this method is that it
can learn from a specific system without using any model simplifications, and it can be
modified easily by re-creating boundary points as the system changes.
Further validation of the KBES Controller could be performed by implementing the
6.3 Future Work 71
controller on a system with fewer simplifications. For example, by imposing a state-of-
charge constraint on the ESS, or by taking into consideration diesel ramp rates, the KBES
can be tested on a more accurate model of the system. Another means would be to
incorporate different RE sources into the microgrid in addition to, or in place of, the
WTG, and observe whether or not the KBES Controller can handle such situations. In
terms of an economical analysis, a financial value for carbon credits [74] can be added to
the evaluation to incorporate the environmental benefits of an ESS. As detailed in [75],
government policies that address socio-economic effects such as social cost, pollution, and
investor return may change the energy dispatch in a complimentary or conflicting manner,
or it may have no effect at all.
As with all aspects of research, there is no ending point to the project as there are many
ways in which the KBES Controller can continue to be ameliorated or be used in many
different microgrid situations.
72
73
References
[1] P. Lautier, M. Prevost, P. Ethier, P. Martel, and L. Lavoie, “Off-grid diesel powerplant efficiency optimization and integration of renewable energy sources,” in ElectricalPower Conference, 2007. EPC 2007. IEEE Canada, Oct. 2007, pp. 274–279.
[2] C. Abbey, J. Robinson, and G. Joos, “Integrating renewable energy sources and storageinto isolated diesel generator supplied electric power systems,” in Power Electronicsand Motion Control Conference, 2008. EPE-PEMC 2008. 13th, 2008, pp. 2178–2183.
[3] F. Katiraei and C. Abbey, “Diesel plant sizing and performance analysis of a remotewind-diesel microgrid,” in Power Engineering Society General Meeting, 2007. IEEE,June 2007, pp. 1–8.
[4] E. Castronuovo and J. Lopes, “On the optimization of the daily operation of a wind-hydro power plant,” IEEE Trans. Power Syst., vol. 19, no. 3, pp. 1599–1606, Aug.2004.
[5] M. Korpas and A. Holen, “Operation planning of hydrogen storage connected to windpower operating in a power market,” IEEE Trans. Energy Convers., vol. 21, no. 3, pp.742–749, Sept. 2006.
[6] M. Korpas, A. T. Holen, and R. Hildrum, “Operation and sizing of energy storage forwind power plants in a market system,” International Journal of Electrical Power &Energy Systems, vol. 25, no. 8, pp. 599 – 606, 2003.
[7] J. Barton and D. Infield, “Energy storage and its use with intermittent renewableenergy,” IEEE Trans. Energy Convers., vol. 19, no. 2, pp. 441–448, June 2004.
[8] J. Carrasco, L. Franquelo, J. Bialasiewicz, E. Galvan, R. Guisado, M. Prats, J. Leon,and N. Moreno-Alfonso, “Power-electronic systems for the grid integration of renewableenergy sources: A survey,” IEEE Trans. Ind. Electron., vol. 53, no. 4, pp. 1002–1016,June 2006.
[9] W. Li and G. Joos, “Comparison of energy storage system technologies and configura-tions in a wind farm,” in IEEE Power Electronics Specialists Conference, 2007., 2007,pp. 1280–1285.
74 References
[10] J. Paatero and P. Lund, “Effect of energy storage on variations in wind power,” WindEnergy, vol. 8, no. 4, 2005.
[11] S. Schoenung and W. Hassenzahl, “Long vs. Short-Term Energy Storage: SensitivityAnalysis,” EPRI and the U.S. Department of Energy, 2007.
[12] C. Abbey and G. Joos, “A stochastic optimization approach to rating of energy storagesystems in wind-diesel isolated grids,” IEEE Trans. Power Syst., vol. 24, no. 1, pp.418–426, Feb. 2009.
[13] M. Ross, R. Hidalgo, and G. Joos, “A knowledge based expert system for the pre-feasibility analysis of an energy storage system in a wind-diesel isolated power grid,”in Power Engineering Society General Meeting, 2009. IEEE, July 2009.
[14] F. Katiraei, R. Iravani, N. Hatziargyriou, and A. Dimeas, “Microgrids management,”in Power and Energy Magazine, IEEE, vol. 6, no. 3, pp. 54–65, May-June 2008.
[15] A.G. Tsikalakis, and N.D. Hatziargyriou, “Centralized Control for Optimizing Micro-grids Operation,” in Energy Conversion, IEEE Transactions on, vol. 23, no. 1, pp.241–248, March 2008.
[16] C.A. Hernandez-Aramburo, T.C. Green, and N.Mugniot, “Fuel consumption mini-mization of a microgrid,” in Industry Applications, IEEE Transactions on, vol. 41,no. 3, pp. 673–681, May-June 2005.
[17] C. Marnay, G. Venkataramanan, M. Stadler, A.S. Siddiqui, R. Firestone, and B. Chan-dran, “Optimal Technology Selection and Operation of Commercial-Building Micro-grids,” in Power Systems, IEEE Transactions on, vol. 23, no. 3, pp. 975–982, August2008.
[18] V. Calderaro, V. Coppola, V. Galdi, V. and P. Siano, “Assessing control and manage-ment issues for micro wind generation integration into microgrids,” in ElectrotechnicalConference, 2008. MELECON 2008. The 14th IEEE Mediterranean, pp. 750-755, May2008.
[19] E. Barklund, N. Pogaku, M. Prodanovic, C. Hernandez-Aramburo, and T.C. Green,“Energy Management in Autonomous Microgrid Using Stability-Constrained DroopControl of Inverters,” in Power Electronics, IEEE Transactions on, vol. 23, no. 5, pp.2346–2352, September 2008.
[20] N.D. Hatziargyriou, A. Dimeas, A.G. Tsikalakis, J.A.P. Lopes, G. Karniotakis, andJ.Oyarzabal, “Management of microgrids in market environment,” in Future PowerSystems, 2005 International Conference on, November 2005.
References 75
[21] B. Billinton, and R. Billinton, “Incorporating well-being considerations in generatingsystems using energy storage,” in Energy Conversion, IEEE Transactions on, vol. 20,no. 1, pp. 225–230, March 2005.
[22] J. McDowall, “High power batteries for utilities - the world’s most powerful batteryand other developments,” in Power Engineering Society General Meeting, 2004. IEEE,pp. 2034–2037, vol. 2, June 2004.
[23] M. Black, and G. Strbac, “Value of Bulk Energy Storage for Managing Wind PowerFluctuations,” in Energy Conversion, IEEE Transactions on, vol. 22, no. 1, pp. 197–205, March 2007.
[24] K. Qian, Z. Li, C. Zhou, and Y. Yuan, “Benefits of energy storage in power sys-tems with high level of intermittent generation,” in Electricity Distribution, 2009 20thInternational Conference and Exhibition on, pp. 1–4, June 2009.
[25] A. Joseph, and M. Shahidehpour, “Battery storage systems in electric power systems,”in Power Engineering Society General Meeting, 2006. IEEE, 2006.
[26] “Fraunhofer UMSICHT, Energy Systems,” 2009. [Online]. Available:http://www.umsicht.fraunhofer.de/englisch/
[27] K. Yoshimoto, T. Nanahara, and G. Koshimizu, “Analysis of data obtained in demon-stration test about battery energy storage system to mitigate output fluctuation ofwind farm,” in Integration of Wide-Scale Renewable Resources Into the Power Deliv-ery System, 2009 CIGRE/IEEE PES Joint Symposium, July 2009.
[28] S.-S. Chen, L. Wang, W.-J. Lee, and Z. Chen, “Power flow control and dampingenhancement of a large wind farm using a superconducting magnetic energy storageunit,” in Renewable Power Generation, IET, vol. 3, no. 1, pp. 23–38, March 2009.
[29] M. Ordonez, M.T. Iqbal, J.E. Quaicoe, and L.M. Lye, “Modeling and optimizationof direct methanol fuel cells using statistical design of experiment methodology,” inElectrical and Computer Engineering, 2006. CCECE ’06. Canadian Conference on,pp. 1120–1124, May 2006.
[30] R. Takahashi, W. Li, T. Murata, and J. Tamura, “An Application of Flywheel En-ergy Storage System for Wind Energy Conversion,” in Power Electronics and DrivesSystems, 2005. PEDS 2005. International Conference on, vol. 2, pp. 932–937, 2005.
[31] A. Shibata, and K. Sato, “Development of vanadium redox flow battery for electricitystorage,” in Power Engineering Journal, vol. 13, no. 3, pp. 130–135, June 1999.
76 References
[32] T.-Y. Lee, “Operating Schedule of Battery Energy Storage System in a Time-of-UseRate Industrial User With Wind Turbine Generators: A Multipass Iteration Parti-cle Swarm Optimization Approach,” in Energy Conversion, IEEE Transactions on,vol. 22, no. 3, pp. 774–782, September 2007.
[33] B.C. Ummels, E. Pelgrum, and W.L. Kling, “Integration of large-scale wind powerand use of energy storage in the netherlands’ electricity supply,” in Renewable PowerGeneration, IET, vol. 2, no. 1, pp. 34–46, March 2008.
[34] S. Schoenung and J. Eyer, “Benefit/cost Framework for Evaluating Modular EnergyStorage,” Sandia National Laboratories and the U.S. Department of Energy, February2008.
[35] O. Shmiel, T. Shmiel, Y. Dagan, and M. Teicher, “Processing of Multichannel Record-ings for Data-Mining Algorithms,” in Biomedical Engineering, IEEE Transactions on,vol. 54, no. 3, pp. 444–453, March 2007.
[36] A. Kusiak, Z. Haiyang, and S. Zhe, “Short-Term Prediction of Wind Farm Power: AData Mining Approach,” in Energy Conversion, IEEE Transactions on, vol. 24, no. 1,pp. 125–136, March 2009.
[37] E.G. Carrano, R.T.N. Cardoso, R.H.C. Takahashi, C.M. Fonseca, and O.M. Neto,“Power distribution network expansion scheduling using dynamic programming geneticalgorithm,” in Generation, Transmission Distribution, IET, vol. 2, no. 3, pp. 444–455,May 2008.
[38] C.-L. Chiang, “Genetic-based algorithm for power economic load dispatch,” in Gen-eration, Transmission Distribution, IET, vol. 1, no. 2, pp. 261–269, March 2007.
[39] R. Aggarwal, and Y. Song, “Artificial neural networks in power systems. III. Examplesof applications in power systems,” in Power Engineering Journal, vol. 12, no. 6, pp.279–287, December 1998.
[40] R. Venkatesan, and B. Balamurugan, “A real-time hardware fault detector using an ar-tificial neural network for distance protection,” in Power Delivery, IEEE Transactionson, vol. 16, no. 1, pp. 75–82, January 2001.
[41] K. Shaban, A. El-Hag, and A. Matveev, “A cascade of artificial neural networks topredict transformers oil parameters,” in Dielectrics and Electrical Insulation, IEEETransactions on, vol. 16, no. 2, pp. 516–523, April 2009.
[42] M.H. Ali, T. Murata, and J. Tamura, “Transient Stability Enhancement by FuzzyLogic- Controlled SMES Considering Coordination With Optimal Reclosing of CircuitBreakers,” in Power Systems, IEEE Transactions on, vol. 23, no. 2, pp. 631–640, May2008.
References 77
[43] S.K. Yee, and J.V. Milanovic, “Fuzzy Logic Controller for Decentralized Stabilizationof Multimachine Power Systems,” in Fuzzy Systems, IEEE Transactions on, vol. 16,no. 4, pp. 971–981, August 2008.
[44] B. Venkatesh, P. Yu, H.B. Gooi, and D. Choling, “Fuzzy MILP Unit CommitmentIncorporating Wind Generators,” in Power Systems, IEEE Transactions on, vol. 23,no. 4, pp. 1738–1746, November 2008.
[45] B. Dos Santos, and V. Mookerjee, “Towards Optimal Expert System Design,” in Pro-ceedings of the Twenty-Fourth Anual Hawaii International Conference on System Sci-ences, vol. 3, pp. 266–273, 1991.
[46] M. Sagheb-Tehram, “Expert Systems Development: Toward Transparency of DesignProcess”, in IEEE International Conference on Information Reuse and Integration,IRI 2007, pp. 257–260, 2007.
[47] L. Galbraith, M. Al-Najjar, and A. Babu, “Expert Systems in Engineering”, in IEEEAES Magazine, pp. 12–14, 1998.
[48] D. Gregory, “Delimiting Expert Systems,” in IEEE Transactions on Systems, Man,and Cybernetics, vol. 16, pp. 834–843, 1986.
[49] P. Fink, and J. Lusth, “Expert Systems and diagnostic expertise in the mechanical andelectrical domain,” in IEEE Transactions on Systems, Man, and Cybernetics, vol. 17,pp. 340–349, 1987.
[50] M. Ross, R. Hidalgo, C. Abbey, and G. Joos, “Energy storage system scheduling foran isolated microgrid,” accepted for publication in Renewable Power Generation, IET,2010.
[51] M. Ross, R. Hidalgo, C. Abbey, and G. Joos, “An expert system for energy storagesystem scheduling in an isolated microgrid,” in Industrial Electronics Society AnnualConference, IECON 2009, IEEE, November 2009.
[52] M. Ross, R. Hidalgo, C. Abbey, and G. Joos, “Analysis of energy storage sizing andtechnologies,” in Electrical Power Energy Conference (EPEC), 2009 IEEE, August2010.
[53] R. Bansal, “Automatic reactive-power control of isolated wind-diesel hybrid powersystems,” Industrial Electronics, IEEE Transactions on, vol. 53, no. 4, pp. 1116–1126,June 2006.
[55] “EW50 Specifications, Entegrity Wind Systems Inc.” 2009. [Online]. Available:http://www.entegritywind.com/pdfs/EW50-Power-Curve.pdf
[56] C. Grigg, P. Wong, P. Albrecht, R. Allan, M. Bhavaraju, R. Billinton, Q. Chen,C. Fong, S. Haddad, S. Kuruganty, W. Li, R. Mukerji, D. Patton, N. Rau, D. Reppen,A. Schneider, M. Shahidehpour, and C. Singh, “The IEEE reliability test system-1996. a report prepared by the reliability test system task force of the applicationof probability methods subcommittee,” IEEE Trans. Power Syst., vol. 14, no. 3, pp.1010–1020, Aug 1999.
[57] A. Kusko, J. DeDad, and N. Exponent, “Short-term, long-term, energy storage meth-ods for standby electric power systems,” in Industry Applications Conference, 2005.Fourtieth IAS Annual Meeting. Conference Record of the 2005, vol. 4, 2005.
[58] M. Coleman, C. K. Lee, C. Zhu, and W. Hurley, “State-of-charge determination fromemf voltage estimation: Using impedance, terminal voltage, and current for lead-acidand lithium-ion batteries,” IEEE Trans. Ind. Electron., vol. 54, no. 5, pp. 2550–2557,Oct. 2007.
[59] S. Lukic, J. Cao, R. Bansal, F. Rodriguez, and A. Emadi, “Energy storage systemsfor automotive applications,” Industrial Electronics, IEEE Transactions on, vol. 55,no. 6, pp. 2258–2267, June 2008.
[60] J. Chawan, C. Abbey, and G. Joos, “VRB Modelling for the Study of Output TerminalVoltages, Internal Losses and Performance,” in Electrical Power Conference, 2007.EPC 2007. IEEE Canada, Oct. 2007, pp. 387–392.
[61] I. Moghram and S. Rahman, “Analysis and evaluation of five short-term load forecast-ing techniques,” IEEE Trans. Power Syst., vol. 4, no. 4, pp. 1484–1491, Nov 1989.
[62] S. Eckroad and I. Gyuk, “EPRI-DOE Handbook of Energy Storage for Transmission& Distribution Applications,” EPRI and the U.S. Department of Energy, 2003.
[63] “Wind data 150 MW Elk River wind plant, Kansas,” 2009. [Online]. Available:http://eece.ksu.edu/ gjohnson/
[65] “Average retail prices for diesel in Canada.” 2008. [Online]. Available:http://fuelfocus.nrcan.gc.ca
[66] C. Abbey, K. Strunz, and G. Joos, “A Knowledge-Based Approach for Control ofTwo-Level Energy Storage for Wind Energy Systems,” in Energy Conversion, IEEETransactions on, vol. 24, no. 2, pp. 539–547, June 2009.
References 79
[67] T. Ise, M. Kita, and A. Taguchi, “A hybrid energy storage with a SMES and secondarybattery,” in Applied Superconductivity, IEEE Transactions on, vol. 15, no. 2, pp. 1915–1918, June 2005.
[68] J. Zeng, B. Zhang, C. Mao, and Y. Wang, “ T. Ise, M. Kita, and A. Taguchi, “Use ofBattery Energy Storage System to Improve the Power Quality and Stability of WindFarms,” in Power System Technology, 2006. PowerCon 2006. International Conferenceon, pp. 1–6, October 2006.
[69] W. Li and G. Joos, “A power electronic interface for a battery supercapacitor hy-brid energy storage system for wind applications,” in Power Electronics SpecialistsConference, 2008. PESC 2008. IEEE, pp. 1280–1285, June 2008.
[70] G. Guidi, T.M. Undeland, and Y. Hon, “An Interface Converter with Reduced VARatings for Battery-Supercapacitor Mixed Systems,” in Power Conversion Conference- Nagoya, 2007. PCC ’07, pp. 936–941, April 2007.
[71] B. Lu, X. Wu, H. Figueroa, and A. Monti, “A low-cost real-time hardware-in- the-loop testing approach of power electronics controls,” in Industrial Electronics, IEEETransactions on, vol. 54, no. 2, pp. 919-931, April 2007.
[72] S. K. Kim, J. H. Jeon, C. H. Cho, J. B. Ahn, and S. H. Kwon, “Dynamic modeling andcontrol of a grid-connected hybrid generation system with versatile power transfer,”in Industrial Electronics, IEEE Transactions on, vol. 55, no. 4, pp. 1677-1688, April2008.
[73] J. Eno, and C.W. Thompson,“Generating Synthetic Data to Match Data Mining Pat-terns,” in Internet Computing, IEEE, vol. 12, no. 3, pp. 78–82, May-June 2008.
[74] I. Falconett, and K. Nagasaka, “A stochastic model to analyze the economic compet-itiveness of wind power projects within a restructured electricity industry,” in PowerEngineering Conference, 2007. IPEC 2007. International, pp. 30–35, December 2007.
[75] S. Wong, K. Bhattacharya, and J.D. Fuller, “Environmental and economic analysis ofOntarios Standard offer Program for small power producers,” in Power and EnergySociety General Meeting - Conversion and Delivery of Electrical Energy in the 21stCentury, 2008 IEEE, pp. 1–7, 2008.