Naval Postgraduate School: Monterey, California

NAVAL
POSTGRADUATE
SCHOOL
MONTEREY, CALIFORNIA
THESIS
MODELING AND OPTIMIZING GREEN MICROGRIDS

AT REMOTE U.S. NAVY ISLANDS
by
Kyle D. Kobold
December 2017
Thesis Advisor: Oleg A. Yakimenko

Second Reader: Fotis A. Papoulias
Approved for public release. Distribution is unlimited.

THIS PAGE INTENTIONALLY LEFT BLANK
REPORT DOCUMENTATION PAGE Form Approved OMB
No. 0704–0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing
instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of
information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions
for reducing this burden, to Washington headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson
Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project
(0704-0188) Washington, DC 20503.
1. AGENCY USE ONLY 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
(Leave blank) December 2017 Master’s thesis
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
MODELING AND OPTIMIZING GREEN MICROGRIDS AT REMOTE U.S.
NAVY ISLANDS
6. AUTHOR(S) Kyle D. Kobold
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING
Naval Postgraduate School ORGANIZATION REPORT
Monterey, CA 93943-5000 NUMBER
9. SPONSORING /MONITORING AGENCY NAME(S) AND 10. SPONSORING /
ADDRESS(ES) MONITORING AGENCY
N/A REPORT NUMBER
11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the
official policy or position of the Department of Defense or the U.S. Government. IRB Protocol number ____N/A____.
12a. DISTRIBUTION / AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
Approved for public release. Distribution is unlimited. A
13. ABSTRACT (maximum 200 words)
This thesis builds upon existing research involving energy microgrid solutions and applies the
findings to isolated U.S. Navy locations, specifically, San Nicolas Island. This includes accurately building
power system models for U.S. Navy remote and disparate island facilities total power-system microgrids
and providing vital information for decision makers. Multifactor optimization and analysis conducted in
this thesis led to enhancement of the proposed models and can aid in development of efficient control
solutions. These solutions would allow the U.S. Navy to efficiently manage power systems at facilities
worldwide. Ultimately, these optimizations can lead to net-zero energy solutions. The experimental and
analytical methods presented in this thesis detail the technical simulations from the EnergyPLAN software
model. By implementing these plans, the models and approaches developed in this thesis can be applied to
other locations, as well.
14. SUBJECT TERMS 15. NUMBER OF

complex system, microgrid, optimization, renewable energy PAGES
75
16. PRICE CODE
17. SECURITY 18. SECURITY 19. SECURITY 20. LIMITATION
CLASSIFICATION OF CLASSIFICATION OF THIS CLASSIFICATION OF ABSTRACT
REPORT PAGE OF ABSTRACT
Unclassified Unclassified Unclassified UU
NSN 7540–01-280-5500 Standard Form 298 (Rev. 2–89)
Prescribed by ANSI Std. 239–18
i
ii
Approved for public release. Distribution is unlimited.
MODELING AND OPTIMIZING GREEN MICROGRIDS AT REMOTE U.S.

NAVY ISLANDS
Kyle D. Kobold
Lieutenant Commander, United States Navy
M.S., University of San Diego, 2010
B.S., University of Georgia, 2003
Submitted in partial fulfillment of the

requirements for the degree of
MASTER OF SCIENCE IN SYSTEMS ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL

December 2017
Approved by: Oleg A. Yakimenko

Thesis Advisor
Fotis A. Papoulias
Second Reader
Ronald E. Giachetti
Chair, Department of Systems Engineering
iii
iv
ABSTRACT
This thesis builds upon existing research involving energy microgrid solutions and
applies the findings to isolated U.S. Navy locations, specifically, San Nicolas Island. This
includes accurately building power system models for U.S. Navy remote and disparate
island facilities total power-system microgrids and providing vital information for
decision makers. Multifactor optimization and analysis conducted in this thesis led to
enhancement of the proposed models and can aid in development of efficient control
solutions. These solutions would allow the U.S. Navy to efficiently manage power
systems at facilities worldwide. Ultimately, these optimizations can lead to net-zero
energy solutions. The experimental and analytical methods presented in this thesis detail
the technical simulations from the EnergyPLAN software model. By implementing these
plans, the models and approaches developed in this thesis can be applied to other
locations, as well.
v
vi
TABLE OF CONTENTS
I. INTRODUCTION..................................................................................................1
A. BACKGROUND AND PROBLEM STATEMENT ...............................1
B. PROPOSED SOLUTIONS .......................................................................4
C. RESEARCH QUESTIONS .......................................................................5
D. RESEARCH APPROACH AND THESIS ORGANIZATION .............5
II. MICROGRIDS AND TOOLS FOR THEIR ANALYSIS ..................................7

A. MICROGRID COMPONENTS ...............................................................7
B. HOMER ......................................................................................................9
C. ENERGYPLAN........................................................................................11
III. SAN NICOLAS ISLAND ENERGY SYSTEM .................................................15

A. ISLAND PROFILE..................................................................................15
B. 1996 – 2017 STUDIES .............................................................................19
1. NREL 1996 Study ........................................................................19
2. NREL 2008 Study ........................................................................20
3. NAVFAC EXWC 2017 HOMER Study .....................................21
C. ARCHITECTURE OF ENERGYPLAN MODELS .............................21
IV. DATA COLLECTION AND MODELING .......................................................23

A. SYSTEMS ENGINEERING ELEMENTS ............................................23
B. DATA COLLECTION AND PROCESSING .......................................24
1. Diesel Gensets ...............................................................................25
2. Wind ..............................................................................................26
3. Solar ..............................................................................................26
4. Cost ................................................................................................28
C. MS EXCEL OPTIMIZATION ...............................................................28
D. ENERGYPLAN REFERENCE MODEL ..............................................33
V. ANALYSIS ...........................................................................................................35
A. REFERENCE MODEL ...........................................................................35
B. OPTIMIZED MODEL ............................................................................37
C. NET-ZERO MODEL...............................................................................39
D. SENSITIVITY ANALYSIS ....................................................................41
VI. CONCLUSIONS AND RECOMMENDATIONS .............................................43

A. CONCLUSIONS ......................................................................................43
vii
B. RECOMMENDATIONS .........................................................................44
APPENDIX A. FINDING AND INPUTTING DATA INTO ENERGYPLAN—

UNITED STATES NAVY (FIDE-USN) ............................................................47
APPENDIX B. DATA PREPROCESSING MATLAB CODE ....................................51
LIST OF REFERENCES ................................................................................................53
INITIAL DISTRIBUTION LIST ...................................................................................55
viii
LIST OF FIGURES
Figure 1. Microgrid Components. Source: Anderson and Yakimenko (2017). ..........7
Figure 2. Typical Energy System in EnergyPLAN. Source: EnergyPLAN

(2017). ........................................................................................................12
Figure 3. California Channel Islands. Source: Google Maps (2017). .......................17
Figure 4. San Nicolas Island Topographical Map. Source: Google Maps (2017).
....................................................................................................................17
Figure 5. Official U.S. Navy SNI Infrastructure Map. Source: NAVFAC

EXWC, Port Hueneme, CA. (2017)...........................................................18
Figure 6. San Nicolas Island Energy Grid. Source: Anderson et al. (2017). ............22
Figure 7. San Nicolas Island OV-1. Adapted from Google Maps (2017). ................24
Figure 8. Direct Normal Irradiance Graph. Source: NSRDB Data Viewer

(2017) .........................................................................................................27
Figure 9. NREL Physical Solar Model (PSM) Data Download Options. Source:
NSRDB Data Viewer (2017). ....................................................................27
Figure 10. Current SNI Energy Structure....................................................................29
Figure 11. Optimized and Net-Zero SNI Energy Structure.........................................30
Figure 12. MS Excel SNI Optimization Interface. ......................................................31
Figure 13. SNI Reference Model Annual Energy Balance. ........................................36
Figure 14. SNI Optimized Model Annual Energy Balance.........................................38
Figure 15. SNI Net-Zero Model Annual Energy Balance. ..........................................40
Figure 16. Sensitivity Analysis for Reference, Optimized, and Net-Zero Energy
Models........................................................................................................42
Figure 17. MATLAB Code Used to Process SNI Energy Demand Data Files.
Source: Yakimenko (2017). .......................................................................51
ix
x
LIST OF TABLES
Table 1. SNI Diesel Gensets. ...................................................................................16
Table 2. Reference, Enhanced, and Net Zero Model Configurations. .....................33
Table 3. Research Questions and Conclusions ........................................................44
xi
xii
LIST OF ACRONYMS AND ABBREVIATIONS
CAES compressed air energy storage
DNI direct normal irradiance
DON Department of the Navy
DOD Department of Defense
ESTEP Energy Systems Technology Evaluation Program
FIDE Finding and Inputting Data into EnergyPLAN
FIDE-USN Finding and Inputting Data into EnergyPLAN – U.S. Navy
HOMER Hybrid Optimization Model for Multiple Energy Resources
INCOSE International Council on Systems Engineering
NPS Naval Postgraduate School
NAVFAC EXWC Naval Facilities Engineering and Expeditionary Warfare Center
NREL National Renewable Energy Laboratory
ONR Office of Naval Research
PESWG Power and Energy Working Group
PSM Physical Solar Model
PV photovoltaics
RE renewable energy
RSWG Resilient Systems Working Group
SE systems engineering
SNI San Nicolas Island
SoS system of systems
xiii
xiv
EXECUTIVE SUMMARY
U.S. Navy personnel operate globally, employ ships across all oceans, and conduct
mission vital to national security on a daily basis. A major piece of successfully doing so
is relying on efficient and effective facilities that are capable of supporting the warfighter
at all times. Many of these locations are disparate and remote, often “islanded” from other
power grids, if not actual islands in the middle of vast oceans. To continually operate these
facilities, requires massive amounts of natural resources for reliable energy production.
Taking advantage of microgrid technology may greatly enhance the U.S. Navy’s energy
security, energy resilience, and overall mission accomplishment.
This thesis produced the following:
 A fully modeled analysis of energy capacity, demand, balancing, and cost

at San Nicolas Island (SNI).
 A step-by-step guide in the form of the FIDE-USN to replicate the

analysis at other locations.
 Various technical simulations and analysis of the SNI total energy

balance, including a net-zero solution using solar and a storage battery.
 Systems-level thinking in terms of the strategic (mission assurance),

technical (efficiencies of microgrid), and financial considerations.
There are multiple tools that are used in this thesis to analyze microgrid technology
and the applicability to SNI. These include EnergyPLAN energy grid management
software, MATLAB, Microsoft Excel, multifactor optimizations, and analytical systems
level thinking and proposals. By comparing the EnergyPLAN results at the Samsø Island
in Denmark, Isle of Eigg in Scotland and also considering the success of renewable energy
at El Hierro in the Canary Islands, it may be possible to enable decision makers to create
the first 100% renewable energy (RE) military facility at SNI. According to Anderson and
Yakimenko (2017, 2), both Samsø and El Hierro are 100% RE, supporting 3,700 and
11,000 people, respectively. Also, there are many similarities between the Isle of Eigg and
xv
SNI, to that point that what has been accomplished on the Isle of Eigg has a great
opportunity to be replicated at SNI. Land mass, max elevation, average temperatures,
average wind speed, and average rainfall are all very comparable.
This thesis adds to two previous National Renewable Energy Laboratory studies
(1996 and 2008) and a recent NAVFAC EXWC study in 2017 regarding energy
management on SNI. The energy opportunity at SNI is both unique and significant. It was
shown that the location, operations, and energy requirements of SNI provide a great
opportunity for the DOD to achieve a tremendous benefit as a first of its kind net-zero
facility at SNI. Figure 1 shows the net-zero model for SNI. Annual load is red, annual
generation for the diesel is black and for the wind turbines is blue, and the total energy
balance is green.
Figure 1. SNI Net-Zero Model Results.
Reference
Anderson, William W., and Oleg A. Yakimenko. “Comparative Analysis of Two

Microgrid Solutions for Island Green Energy Supply Sustainability.” IEEE
International Conference on Renewable Energy Research and Applications. San
Diego, CA, 2017.
xvi
ACKNOWLEDGMENTS
To my wife, Devon, and our 19-month-old daughter, Alexandria May, thank you
for your unwavering love and support during this time, which has allowed me to devote an
all-out effort to this thesis.
xvii
xviii
I. INTRODUCTION
A. BACKGROUND AND PROBLEM STATEMENT
Microgrids are small-scale power systems located closer to the load than typically
found in conventional power plants. A traditional microgrid typically includes three core
components: hybrid generation, energy storage, and controls. All of these components
work together as a system solution to serve a nearby load, such as a wind turbine and a
storage battery. The key difference with green microgrids is that they leverage an
alternative energy source in the power generation phase. These sources can include wind
(on shore or off shore), solar, hydrodynamic, pumped hydrodynamic, river, wave,
compressed air energy storage (CAES), and a number of other emerging technologies.
The processes required of the U.S. Navy to procure, move, store, and utilize various
forms of energy are fraught with security vulnerabilities and massive costs. Simply put, the
U.S. Navy spends much human and monetary capital on supplying its ever-growing energy
demands and physical footprints around the world. This essential national security resource
is in need of optimization and simplification in the way that it is used by our military,
especially in isolated and disparate island facilities. As a specific example, energy usage
on San Nicolas Island (SNI) will be modeled and analyzed for optimization in this thesis.
The ability to secure the energy supply and decrease vulnerabilities is of high
interest and crucial in the ever-changing world. Threats from ill-intentioned actors are
continually increasing and costs due to manpower and transportation play a hefty role in
the bottom line. Traditionally, our government has to coordinate complex shipments of oil
and gas to the far reaches of the world where our military has a presence. These island
locales are responsible not only for their own needs for energy on island, but for the
resupply and support of our nation’s Navy when present.
The International Council on Systems Engineering (INCOSE) believes that power,

energy, and resilient systems are of such high importance that they currently have two
working groups that focus their efforts on improving the body of knowledge in these
globally key, vital areas. These working groups are the Power and Energy Working Group
1
(PESWG) and the Resilient Systems Working Group (RSWG). The INCOSE working
groups’ efforts are directly applicable to the proposals in this thesis regarding remote island
locations under control of the U.S. Navy. These critical locations can definitely benefit
from improvements and optimizations in the area of resilient power and energy systems.
According to INCOSE (2017), the PESWG seeks to organize experts from within
the ranks of INCOSE as well as other professionals in the energy sector of the economy to
facilitate a “systems approach” to the analysis and future development of effective energy
solutions. The purpose of the PESWG is to direct expertise and a “systems” focus to
support decision makers in the critical challenges of developing future energy systems.
This is directly applicable to the fields of energy and microgrids in terms of their
application to the U.S. Navy.
The use of microgrids with varying combinations of traditional and renewable

energy sources on the grids can deliver these results for the U.S. Navy. Microgrids are
small-scaled power systems located closer to the load (item using the power) than is
typically found in conventional power plants. A microgrid normally includes three core
components: hybrid generation, energy storage, and controls. All of these components
work together as a system solution to serve a nearby load. Green microgrids leverage an
alternative energy source in the power generation.
For the U.S. Navy, achieving an outlook similar to this for energy systems can lead
to several overall enhancements:
 all aspects of energy security, especially at remote and disparate island

facilities
 energy system resiliency against threats
 cost savings from energy logistics
 cost savings from decreased manpower requirements
 scalability as needed for other remote and isolated U.S. Navy installations
worldwide
2
In his thesis regarding system of systems (SoS) life cycle security architecture and
design, Harley (2016) stated that with the increasing complexity of systems and the
attacker’s improving ability to conduct multiple unpredictable, non-linear, multi-domain
attacks, the ability of any system to survive and accomplish its designed objectives greatly
depends upon its security quality. Transitioning the U.S. Navy’s most remote and
vulnerable key locations to a more energy-secure posture can greatly enhance total fleet
readiness and reliability.
This thesis contributes to the existing research of optimizing renewable energy

microgrids to enhance energy security for remote island naval installations. The U.S. Navy
is a traditional energy consumer, on a massive scale. The requirement to operate
worldwide, all days of year, and all hours of the day places enormous demands on the
amount and types of energy required to sustain the force. Efficiency in costs is not at the
forefront of concerns as much as availability and capability.
Disparate locations often require extenuating circumstances and heavy demands to

be able to continually man and operate those forces necessary to support the U.S. Navy in
a continually sustained manner around the world. Vulnerabilities exist all along the supply
chain, the transportation process, the manpower footprint, and the overall total energy
system. These facilities and critical operations centers are increasingly vulnerable to power
disruptions such as power outages. Thus, a dependable and self-sufficient energy solution
is of the utmost importance. Currently, critical facilities heavily rely upon diesel generators.
Challenges to this stated problem include ad-hoc micro grid integration, grid stability
management, developing modular approach, and battery technology selections.
This thesis suggests the U.S. Navy can benefit from transitioning to advanced
energy system microgrid technology while maximizing use of renewable resources as an
energy supply. This is especially applicable and vital at the most isolated locations where
the mentioned risks to the system are inherently higher than in a more protected location
on the mainland in the continental United States.
3
B. PROPOSED SOLUTIONS
The Department of the Navy (DON) needs to increase energy security to ensure
mission assurance of critical infrastructure primarily through microgrids at remote islanded
sites. This thesis intends to provide four contributions to the fields of systems engineering
and energy. The PESWG and RSWG may also benefit from the findings presented here.
Specifically, the following will be provided:
 a fully modeled analysis, as a SoS, of installed energy capacity, demand,

balancing, and cost at SNI that can be replicated for other similar U.S.
Navy locations
 technical simulations and analysis of the SNI total energy balance
 technical multifactor optimizations of the generation and storage models
 microgrid architecture design analysis and proposals to enhance future

microgrid energy security and resiliency at U.S. Navy isolated island
locations
 systems-level thinking in terms of the strategic (mission assurance),

technical (efficiencies of microgrid), and financial considerations. This is
one step beyond the optimizations and considers the holistic factors that
should be integrated into the systems engineering (SE) design approach
More importantly, the models, analyses, and optimizations presented herein can be
replicated at other U.S. Navy island locales. Appendix A details the step-by-step process
used here that can easily be applied at other installations. This thesis utilizes raw data
collected from installed equipment at SNI and analytical simulations run by EnergyPLAN
energy software.
Renewable energy can be a major advantage in terms of energy security, resiliency

to vulnerabilities, costs, and manpower requirements. There is much opportunity for the
U.S. Navy to apply microgrid architecture to all existing applicable technology at each of
its isolated locations to improve overall combat effectiveness by enhancing energy security
4
and resiliency. Microgrids are small-scale power systems located closer to the load than
those typically found in conventional power plants. Some small-scale power systems,
specifically green microgrids, leverage an alternative energy source in the power
generation process.
The recommendations presented here suggest viewing an island microgrid energy

system as an SoS that can all be monitored, controlled, operated through an internet of
things approach. The Department of the Navy needs to increase energy security to ensure
mission assurance of critical infrastructure primarily through microgrids at remote islanded
sites.
C. RESEARCH QUESTIONS
This thesis aims to investigate four key questions:
 Is it possible to accurately model and apply power system microgrid

technology at U.S. Navy remote and disparate island facilities?
 Can the derived model lead to a more efficient total system energy balance
for San Nicolas Island and other U.S. Navy islands?
 Can multifactor optimizations enhance the proposed models and lead to

enhanced control solutions for the power system microgrids?
 Can U.S. Navy bases become more energy self-sufficient to the point of
increasing energy security and resiliency?
D. RESEARCH APPROACH AND THESIS ORGANIZATION
The work presented in this thesis builds upon existing research and ongoing
analysis of microgrid solutions for island green energy supply and sustainability.
Experimental and analytical SE approaches were used to create an EnergyPLAN model for
SNI. Data was collected from the existing SNI energy grid and modeled in the
EnergyPLAN software to compare all existing energy aspects and optimize the results.
Technical simulations were also conducted and analyzed.
5
The Academic Associate Energy Academic Group seminars and the Office of
Naval Research (ONR) Energy Systems Technology Evaluation Program (ESTEP)
presentations were both utilized to the full extent possible to contribute to this thesis. Both
events were conducted on campus at the Naval Postgraduate School (NPS) in June and July
2017. Energy and academic research experts and professionals from across the country
presented a varying field of topics, many of which are applicable to the subject matter in
this thesis. Also, a site survey of SNI was performed in September 2017 with Naval
Facilities Engineering and Expeditionary Warfare Center (NAVFAC EXWC) personnel to
fully understand and capture the aspects of the existing energy infrastructure at SNI. This
greatly enhanced the energy model in EnergyPLAN.
This thesis is organized as follows. Chapter II introduces microgrids and tools for
their analysis, such as HOMER and EnergyPLAN. Chapter III describes the details of the
complete SNI energy system. Chapter IV contains the data collection process and modeling
efforts. Chapter V contains the modeling results, conclusions, and recommendations.
6
II. MICROGRIDS AND TOOLS FOR THEIR ANALYSIS
This chapter introduces microgrids, their architectures and components, and

provides a brief review of software packages most applicable to microgrid analysis and
optimization.
A. MICROGRID COMPONENTS
Microgrids are directly applicable to the U.S. Navy’s interest to optimize whole
system energy plants at isolated locations. Through doing so, the U.S. Navy can become
more energy secure, especially at remote and disparate island facilities. Utilizing optimized
green microgrids can also lead to energy system resiliency against threats and cost savings
from energy logistics and decreased manpower.
A microgrid, as shown in Figure 1, is a small-scaled power systems located closer

to the load than typically found in conventional power plants. A microgrid normally
includes three core components: hybrid generation, energy storage, and controls.
Figure 1. Microgrid Components. Source: Anderson and Yakimenko (2017).
All of these components work together as a system solution to serve a nearby load.
Most microgrids are designed and installed to meet a specialized need not ideally served
by the utility company. Often this need is dictated by the remoteness and dislocation of the
load from a utility company such as a remote island or by loads that are deemed critical
infrastructure. For remote island communities, the microgrids have been used to provide
greater independence, reliability and sustainability from off-island power services. As a
result, these green microgrids have rather creative and complex designs.
7
Microgrid deployment relies on fast and real time control, adequate capacity and
robust communications. Modular, building level controls can alleviate information
assurance issues of information technology interconnections. Modular power control can
be sized to meet specific needs. With power storage, available renewables can be fully
utilized. With correctly sized, dispatchable generation, seamless grid to off-grid operation
for power resiliency is possible. Though ideal for mission critical facilities, adoption may
be slow. However, for mission essential facilities application of non-dispatchable power
sources can be utilized to form a microgrid and limit the dependency on backup generation.
These facilities can provide an intelligent, scalable, microgrid power system designed to
distribute, manage and control AC and DC power generated and consumed in a grid
connected or stand-alone environment.
For U.S. Navy installations, microgrids have been used and considered primarily
to serve critical infrastructure typically employed in remote locations to include SNI.
Military installations certainly have very specific energy needs that also vary slightly from
location to location. The U.S. Department of Energy classifies a microgrid as a complex
system, not so much for its characteristics of emergent behavior or nonlinear dynamics,
rather simply because of its use of advanced distributed energy resources components.
According to Sugiyama (2017), microgrids can include
 small, local and stand-alone power systems integrated with a larger

distribution feeder
 energy storage and distributed generation (DG) within a small control area
 variety of DG
 plug-and-play functionality not dependent upon communications
 single interface to power system to have seamless power transition

between parallel to grid and islanded
Although there has been a steady push into architecting “smart microgrids” by
optimizing the entire system, this approach has not effectively leveraged complex systems
8
tools, such as agent-based modeling. However, for remote islands and U.S. Navy
installations the primary benefit is energy security and mission assurance, not economic
(Anderson and Yakimenko, 2017). A microgrid should be interconnected to the
installation’s grid as well as be considered interconnected as a networked system with other
systems. This leads to optimization of control systems to factor in the dynamics of weather,
utilities commodity pricing, and real-time loads.
Microgrids consist of parts that are numerous, diverse, interdependent, connected,

and adaptive. They deal with systems that are not normal in that the probabilities of
outcomes are not described by normal distributions. Power law distributions are the norm.
Given the variability of renewable energy generation serving small and disparate loads
coupled with the system operation of a microgrid, these microgrids can conceivably be
considered a complex system by virtue of their interrelated, heterogeneous elements
(agents and objects) (Anderson and Yakimenko 2017). Control Systems are used in
microgrids to primarily match generation to the load. Control systems for microgrids vary
from simple inverter controls to more complex controls systems using programmable logic
controllers.
B. HOMER
The NREL hybrid optimization model for multiple energy resources (HOMER) has
been the gold standard for energy grid analysis and optimizations. It is used worldwide and
has been very successful. However, using EnergyPLAN to model microgrids, especially
green microgrids, has distinct advantages over HOMER in the following areas.
 EnergyPLAN is free. There is free account access, free software

download, free reference models from around the world, and a free
discussion board and forum or share ideas and learn about issues.
 The user interface is designed as a series of tab sheets and stacked side
columns. Therefore, jumping between sections and inputting data is very
quick and easy. Also, there is online training available from the
EnergyPLAN website.
9
 Renewable energy systems, such as wind energy, tend to fluctuate greatly
throughout any measured time period. Connolly (2015, 4) states that
EnergyPLAN considers the three primary sectors of an energy system to
be electricity, heat, and transport, integration of these fluctuating sectors
becomes more of an issue. This is even more pronounced when these
renewable sources come to achieve more penetration in the grid.
EnergyPLAN allows all of this to occur.
 EnergyPLAN has been used to simulate a 100% renewable energy system

for Denmark. (Anderson and Yakimenko 2017, 4)
 HOMER does not account for transients of equipment and can lead to the
output showing certain pieces of equipment, such as a diesel generating set
(genset), being switched on and off more often than may be realistic. This
can lead to errors in the outputs.
 EnergyPLAN, when compared to many other energy-planning tools,

includes the ability to add a plethora of renewable energy options to the
traditional energy grid. The result is a fully analyzed, both technical and
economic, hybrid microgrid for decision makers to choose the best course
of action.
 HOMER will always optimize for cost first, not the best technical
solution.
 EnergyPLAN, while offering the option to maximize the technical

solution, can also simulate the costs of an energy system in four areas as
stated by Lund and Connolly:
 Fuel costs: purchasing, handling, and taxes for each fuel as well as
CO2 costs.
 Investment costs: capital required, the lifetime of each unit, and the
interest rate on repayments.
10
 Operation costs: the variable and fixed operation and maintenance
costs for each production unit.
 Additional costs: any extra costs not already accounted for, such as
the cost of insulating houses for increased energy efficiency.
(2017, 28)
C. ENERGYPLAN
The computer software EnergyPLAN: Advanced Energy Systems Analysis

Computer Model version 12.0 was designed by Henrik Lund (2016) at the Sustainable
Energy Planning Research Group at Aalborg University in Denmark. It is intended to
simulate energy systems, specifically green microgrids, by taking user inputs and running
selected simulations on an energy system. The overall structure consists of demand, supply,
balancing and storage, cost, simulation, and output. The user inputs control the demand,
supply, balancing and storage, cost, and desired selection of simulation options. Major
components of the user inputs include supply data, demand data, renewable energy sources,
energy plant capacities, and costs. Outputs include energy balances, annual productions,
fuel consumption, and total costs. Using an SE approach and viewing a microgrid as a SoS,
it assists the “design of energy planning strategies based on technical and economic
analyses of the consequences of different energy system options and investments”
(EnergyPLAN 2017). This system has thus far been most directly applicable to European
nations but is now, for the sole purpose of this thesis, being analyzed for use at U.S. Navy
installations. It is a deterministic, hour-simulation model, aggregated in a systems
description through optimizing operations and using analytical programming. The
simulations provide an opportunity to include technical simulations and market-economic
simulations. For these reasons, it appears to be a good fit for research regarding energy
usage at disparate and remote U.S. Navy facilities.
Figure 2 shows a typical energy system block diagram in EnergyPLAN. The overall
structure consists of demand, supply, balancing and storage, cost, simulation, and output.
The user input controls the demand, supply, balancing and storage, cost, and desired selection
of simulation options. Major components of the user inputs include renewable energy sources
11
and energy plant capacities. Outputs include energy balances, annual productions, fuel
consumption, and total costs. The advantages to using EnergyPLAN, when compared to
many other energy-planning tools, include the ability to add a plethora of renewable energy
options to the traditional energy grid. The result is a fully analyzed, both technical and
economic, hybrid microgrid for decision makers to choose the best course of action.
EnergyPLAN has already been used to simulate a 100% renewable energy system on Samsø
Island in Denmark. (Anderson and Yakimenko 2017, 4)
Figure 2. Typical Energy System in EnergyPLAN.

Source: EnergyPLAN (2017).
Through the path of his doctoral research, David Connolly (2015) of Aalborg
University created Finding and Inputting Data into EnergyPLAN (the FIDE Guide). As
the data collection process can be time consuming and difficult, the FIDE paper was relied
upon heavily to form the basis for my SNI data collection strategy. One contribution of this
thesis is the creation of a Finding and Inputting Data into EnergyPLAN for the U. S. Navy
12
(FIDE-USN) by using the original FIDE as a guide along with my research. This can be
found in Appendix A. While EnergyPLAN can be simple to use, the reference model
requires a focused effort to find and input the correct data. Once the set of reference data
is obtained, the model can be adjusted to account for any U.S. Navy system with varying
combinations of energy generation, loads, and storage.
13
14
III. SAN NICOLAS ISLAND ENERGY SYSTEM
This chapter describes in detail all aspects of the existing SNI energy system and
past energy system reviews conducted by the National Renewable Energy Lab (NREL).
A. ISLAND PROFILE
San Nicolas Island is one of the eight islands that comprise the California Channel
Islands, as shown in Figure 3. The others include Anacapa, San Clemente, Santa Barbara, Santa
Catalina, San Miguel, Santa Cruz, and Santa Rosa. The total land mass is approximately 54
square miles (14,500 acres) with the highest peak being 905 feet as shown in the topographical
map in Figure 4. Temperatures range from 56 F to 69 F with annual rainfall just over 7.9”
(Weather Underground 2017). Winds on SNI are from the northwest and strong most of the
year. The average wind speed is 7.2 m/s (14 knots) and only varies slightly throughout the year
(Weather Underground 2017). NREL (1996, 2) states that, during the windiest months, March
through July, wind speeds average 8.2 m/s (16 knots), whereas during the least windy months,
August through February, wind speeds average 6.2 m/s (12 knots). No freezing temperatures
have ever been recorded on SNI.
San Nicolas Island is the most remote and farthest from the California coast of the
group, 61 miles, and is solely used for U.S. Navy purposes. Figure 5 details the government
and military footprint and installations on the island.
San Nicolas Island is not served by a mainland utility company and the only
existing energy infrastructure and generation at SNI includes five diesel generators of
varying capacities and seven 100kW wind turbines. The diesel genset specifics are shown
in Table 1. They typically run at only 30% of nominal capacity, while they are most
efficient at 70% (Anderson et al. 2017). The wind turbines are all Northwind 100 kW
models and waste much of their capacity as heat in a load bank due to a lack of storage
capability on island.
15
Table 1. SNI Diesel Gensets.
Unit (year installed) Rated Capacity Type
diesel genset #1 (1990) 750 kW CAT 3512
diesel genset #2 (2002) 1250 kW Cummins QSK-45
diesel genset #3 (2002) 750 kW Cummins QST-30
diesel genset #4 (?) 900 kW Cummins QST-30
diesel genset #5 (1960) 1000 kW EMD 645
The diesel gensets use JP-5 fuel and produce the vast majority of the power for the
island. The fuel is barged in and pumped up from the beach to the airfield, then up to the
generating plant. As for the power plant infrastructure, the building was constructed in
1960 and expanded in 1990. The distribution panel was installed in 1988. The average load
ranges from 550 kW to 950 kW with a 150 kW to 200 kW peak over the average occurring
in the morning, noon and evening (Anderson et al. 2017, 1). The power factor is currently
0.7-0.75. Power is distributed around the island by 4160 V distribution system, much of it
underground. There are also 32 emergency back-up diesel generators on buildings and
range operations.
From both the NREL (1996 and 2008) studies, the Defense Energy Supply Center
(DESC) covers the cost to transport JP-5 fuel to SNI. As of October 1, 2017, the cost to
purchase JP-5 from DESC is $2.18/gallon (Roth 2017, 2). According to NREL (2008, 20),
barge shipments costs $40,000 each, are able to transport 250,000 gallons of JP-5 fuel, and
are run on average three and a half times per year. These barges must also be run alongside
a standby ship, which costs $15,000 per barge trip.
The SNI transportation fleet is comprised of trucks, vans, and a bus. All vehicles
on the island operate on either gasoline or JP-5. According to NREL (2008, 21), the annual
gasoline demand on the island is 32,000 gallons. This is shipped on a different vessel than
the JP-5 fuel. Regular unleaded gas costs $2.10/gallon to purchase from DESC (Roth 2017,
2), and costs an additional $1.50/gallon to transport to the island. This shipment cost is not
covered by DESC.
16
Figure 3. California Channel Islands. Source: Google Maps (2017).
Figure 4. San Nicolas Island Topographical Map. Source: Google Maps

(2017).
17
Figure 5. Official U.S. Navy SNI Infrastructure Map. Source: NAVFAC EXWC, Port Hueneme, CA. (2017).
18
B. 1996–2017 STUDIES
1. NREL 1996 Study
In 1996, NREL conducted the first of two energy studies regarding energy usage at
SNI. This first study was designed to determine if wind energy could be used to decrease
overall energy costs on the island. The goal was to institute a wind-diesel hybrid system.
At the time, the major energy cost drivers on SNI were capital and installation costs, diesel
fuel costs, and diesel operating and maintenance costs. Due to the generally favorable
environmental conditions to use wind energy at SNI, this goal had a solid chance of success
to deliver the proper amount of energy and to be cost effective. Any excess generation
would not be wasted but could instead be used for reverse osmosis (RO) units, water
heating, or space heating. The major hurdle to the excess energy utilization was the lack of
an existing storage capability in 1996. As a result of the study, other, more detailed,
analyses were recommended. Specifically, dynamic load management was recommended
using load and wind data at shorter intervals than the 1-hour average used in the study (less
than or equal to two minutes recommended).
According to Timothy Olsen and Ed McKenna of NREL (1996, 14), the

assumptions made in 1996 included the following.
First, at least 200 kW must always be generated by the existing diesel

generators even if there is excess wind capacity. Second, it is assumed that
only the necessary number of turbines will be generating power at any given
time, with the remaining turbines idled. Third, a minimum diesel time is
required to hold the number of diesel starts on the order of 100.
The 1996 proposed solution, stated by Olsen and McKenna (1996, 22) included a
hybrid system of diesel and wind.
Between one and four commercially available wind turbines (each with a
capacity of 225 kW) would be combined with the existing 3500 kW diesel
generation capacity. Of note, SNI would later add additional diesel genset
capacity to 4650 kW, which is where it currently stands. With a demand
peak of 1230 kW, no more than 1500 kW of diesel is on line at any time.
Therefore, wind penetration of on-line capacity with four wind turbines is
900/1500 = 60%. Based on instantaneous power, wind penetration can
range from 0% when there is no wind to 250% when peak wind power of
900 kW is combined with a minimum load of 360 kW.
19
Power storage, photovoltaic generation, dump load, and advanced load
management were not included in the 1996 analysis.
The wind and diesel hybrid system produces favorable energy and economic
results. The Olsen and McKenna (1996, 27) NREL report concludes, “the hybrid case using
four 225 kW wind turbines is $0.338/kWh versus $0.358/kWh for the baseline case. This
would create a savings of 5.6%. The payback period is 6.97 years, the internal rate of return
of 13.1%”.
2. NREL 2008 Study
Still interested in how renewable energy may enhance the Department of Defense
(DOD) mission, Alicen Kandt and Andy Walker of NREL (2008) developed the method
used in this study to calculate a hybrid energy system aimed at minimizing life cycle costs.
This renewable assessment and optimization was done to support this determination and to
lay a foundation for the planning of a future “Net Zero Energy Installation.” The DOD
funded this study to provide: “(1) a plan outline for transitioning Navy Outlying Landing
Field SNI into a renewable community; and, (2) a current energy, water and waste baseline
of SNI” (NREL 2008, 6). To take the 1996 study a few steps further, the team collected
data for a baseline of energy, water, and waste.
As stated by the U.S. Energy Information Administration, the Energy Independence

and Security Act, passed in December 2007, defines zero-net energy buildings as “a
building that is designed, constructed and operated to
 require a greatly reduced quantity of energy to operate
 meet the balance of energy needs from sources of energy that do not
produce greenhouse gases
 result in no net emissions of greenhouse gases
 are economically viable” (2017)
Proving that a net-zero facility is possible at SNI is one of the main goals of this
thesis. Additionally, according to NREL (2008, 7),
20
an optimization analysis was done to determine the combination of
renewable energy technologies that could replace the use of JP-5 fuel to heat
and power the stationary loads at San Nicolas Island, first with batteries and
then without batteries. The technologies evaluated include photovoltaics,
wind power, solar ventilation air preheating, solar water heating, solar
thermal steam and solar thermal electric, biomass thermal steam and
biomass electric, and daylighting. Results indicate that solar water heating,
solar ventilation air preheating and daylighting can be integrated directly
into buildings to reduce both electric and heating use. The remaining
electrical requirements can be reduced by wind power connected to the
central plant. Remaining heat uses at some buildings would continue to be
supplied by oil since there is no renewable energy technology considered
that can serve those loads. However, this remaining load could easily be
served by bio-diesel purchased and delivered in the same way that JP-5 is
currently delivered. Also, these oil-fired loads could be converted to
electricity which could be provided by renewables, although this would not
reduce life cycle cost under current conditions.
3. NAVFAC EXWC 2017 HOMER Study
As recently as earlier this year, the SNI microgrid was modelled using HOMER
simulation software. According to Anderson et al. (2017, 1),
various technologies were investigated as potential solutions to increase RE

penetration and resiliency of the SNI microgrid, while reducing reliance on
costly diesel generators. These technologies include solar, wind, battery,
fuel cell, flywheel, and supercapacitor. The results show a solar and battery
installation to be the most cost effective method to increase renewable
penetration, reduce diesel reliance, and increase resiliency of the SNI
microgrid at minimal cost.
This study concluded that there were two worthy configurations that deserved
further analysis. First is the solar with battery storage option. Second, and the most efficient
net-zero solution, is the solar with fuel cell configuration (Anderson et al. 2017, 10).
C. ARCHITECTURE OF ENERGYPLAN MODELS
Based upon the results and conclusions of the 1996 and 2008 NREL studies, as well
as the 2017 NAVFAC EXWC HOMER study of SNI, the following EnergyPLAN models
were created and run to serve as a comparison for further analysis:
Reference (current SNI configuration) model: This model serves as the SNI
reference model, to which all below models will be compared. The SNI EnergyPLAN
21
reference model shows the island’s energy microgrid as it is currently established on site.
The diesel gensets carry the vast majority of the total island load and operate at very
inefficient frequencies due to the constant nature of cycling them up and down. As shown
in the reference model section of Appendix B, the island’s daily load remains fairly
constant throughout the year, averaging about 500 kW daily. One glance at the production
graphs reveals that there is a great opportunity not only to decrease diesel genset
production, but also to capture, store, and use wind turbine energy production on demand.
Thus, a better model must be possible.
The nominal SNI energy microgrid is depicted in Figure 6. The five diesel gensets
and seven wind turbines can be seen, connected with switchgear.
Figure 6. San Nicolas Island Energy Grid. Source: Anderson et al. (2017).
Optimized model. This model adds a storage battery to the base model, decreases
the contribution of the diesel gensets to the SNI energy grid, and increases the RE
penetration.
Net-Zero model. This model adds solar PV and a storage battery to the base model
and shows that SNI can become a 100% RE U.S. Navy installation.
22
IV. DATA COLLECTION AND MODELING
This chapter describes the data collection process, the preprocessing that was
required for the data, and the energy system modeling efforts.
A. SYSTEMS ENGINEERING ELEMENTS
Scott (2016, 4) states that insight is the goal of Law #3 in the “9 Laws of Effective
Systems Engineering.” Through the data obtained from SNI, many insights can be formed
about energy usage and opportunities for improvement on the island. Good information
feeding a good process leads to insight, and insight leads to better choices. Otherwise,
“garbage in yields garbage out.” This is the one of the powers of applying systems
engineering principles to complex systems. Models allow us to capture and communicate
our understanding unambiguously. According to Scott, “models enable us to coherently
reason about the problem and the solution in a way that is not possible in the abstract”
(2016, 4). The applicability of this research to the field of systems engineering is in the
research and solutions to power systems for mission assurance of critical infrastructure.
This is accomplished by expanding the system boundary to best ensure SE tools create
more value through greater efficiency, resiliency, and mission assurance.
Other classic SE process elements that are presented in this thesis include an
operational concept, an effective need, system boundaries, an input-output model, a
functional architecture design, a physical architecture design, and systems analysis
methodology. Figure 7 depicts the SNI energy ecosystem in a high-level operational
concept graphic, or an OV-1.
23
Figure 7. San Nicolas Island OV-1. Adapted from Google Maps (2017).
As shown, SNI currently receives energy from five diesel gensets and seven 100kW
Northwind wind turbines. Most realistic short-term energy solutions, including those plans
that can result in a net-zero energy solution on the island, add in solar photovoltaics (PV)
and storage batteries. Small island usage of storage batteries has already proven to be
successful with Tesla’s venture to American Samoa (Science Alert 2016).
B. DATA COLLECTION AND PROCESSING
In terms of data collection, three sets of data were obtained to create the SNI
reference model. These sets include diesel generator production (island demand), wind
turbine production, and solar PV. The solar PV data was needed in order to analyze
enhanced models, so direct normal solar irradiance (DNI) data was obtained. Monetary
data, such as investments, operations, fuel, and maintenance costs were estimated based on
the three previous reports conducted on SNI discussed in Chapter III. LCDR Kyle Kobold
and Professor Oleg Yakimenko of the Naval Postgraduate School (NPS) visited NAVFAC
EXWC in Port Hueneme, CA on September 6, 2017. They were accompanied on the site
visit by Bill Anderson of NAVFAC EXWC. During the visit, they surveyed the facilities
24
and downloaded actual island demand data from a standalone computer that has access to
the energy system on the island.
1. Diesel Gensets
The three sets of annual SNI demand data were downloaded and exported to a text
file for processing. MATLAB code was written to filter out every hourly data point from
the three files. Since EnergyPLAN requires 8784 data points for the distribution files that
are called in the software, the data processing was a vital step to the creation of the model.
Data preprocessing includes:
 Raw data was obtained from the storage medium of the desired energy
source. Diesel generator demand data was collected on site at Port
Hueneme at NAVFAC EXWC. Wind turbine data was obtained on SNI.
Solar PV data was downloaded from NREL. Existing energy capacities
were obtained from past NREL reports.
 Diesel generator demand data required the most effort to process. Three
separate files, containing 10-minute data for a year were obtained at
NAVFAC EXWC. In order to feed EnergyPLAN the required distribution,
this data needed to be filter to hourly data for a year, or 8784 data points.
MATLAB code was written to extract the data points in this manner. A
sample of the MATLAB code is provided in Appendix A.
 Wind data did not require as much work as it arrived in larger time
intervals and was filtered to hourly data using MS Excel.
 Solar PV data was simply downloaded from NREL and average direct
normal irradiance. From there, the sunrise and sunset times were
calculated and used to arrive at an average hourly value for solar PV
capability at SNI. Since SNI currently does not have solar PV installed,
this is an estimated value.
25
Obtaining raw data and converting it into usable data for EnergyPLAN can require
a considerable amount of effort. Preprocessing of the data is a vital step in the process
though if realistic results are desired.
2. Wind
Wind data was provided by Nicolas DeMarco, one of the authors of the NAVFAC
EXWC HOMER study that was recently conducted on SNI. This data must be acquired on
island, as there is no way to access it remotely. Weather prohibited a trip to the island for
the purposes of this thesis. For accuracy, this data was compared to downloadable wind
speed data from a couple websites – NREL and Weather Underground. As with the diesel
generator data, 8784 data points that represent a full year of hourly data were obtained.
This also required filtering due to arriving as 10-minute wind speed data. The index
method, described in Appendix A, was used to create this distribution file for
EnergyPLAN. That is, the fastest recorded wind speed was set to one, and the rest of the
data points were referenced from there.
3. Solar
Solar PV data was obtained from the NREL Geospatial Data Science website. Once
on the site, one can either select or query a specific data point from the graph, shown in
Figure 8. Or, one can download data by filling in the requested information as shown in
Figure 9 in the Download Wizard. This provides National Solar Radiation Database
(NSRDB) information.
26
Figure 8. Direct Normal Irradiance Graph. Source: NSRDB Data Viewer
(2017)
Figure 9. NREL Physical Solar Model (PSM) Data Download Options.

Source: NSRDB Data Viewer (2017).
According to NREL (2017), this data provides monthly average and annual average
daily total solar resource averaged over surface cells of 0.038 degrees in both latitude and
27
longitude, or nominally 4 km in size. The solar radiation values represent the resource
available to solar energy systems. The data was created using cloud properties which are
generated using multiple NREL proprietary models. The data are averaged from hourly
model output over 17 years (1998-2014).
4. Cost
Information regarding the cost inputs proved difficult to obtain but was found
scattered throughout the three previous detailed reports regarding SNI’s energy system.
For the energy prices used in the multifactor MS Excel optimization model, the Roth (2017,
2) OSD Memo was referenced.
C. MS EXCEL OPTIMIZATION
Since EnergyPLAN requires user inputs to derive the output, it is vital to feed
EnergyPLAN accurate information. Otherwise, the output will be worthless. So, the first
step to creating the most accurate SNI output model as possible is to derive the actual
required numbers of the different types of energy sources on the island. This was
accomplished with MS Excel.
MS Excel optimizations were used to arrive at the specific numbers used in the
optimized and net-zero models found in Chapter V. The optimization of the microgrid
components relied on a formulation that is multi-objective, mixed-integer, and nonlinear.
The feasibility of the recommended configurations was established through constraints
governing the requirements for the load and the capabilities of both the generation, storage
and overall SNI energy system. There are two objective functions that maximize
penetration of renewables onto grid and minimize overall costs. Figure 10 shows the SNI
energy architecture as currently installed while Figure 11 shows an example of an enhanced
model with solar PV and a storage battery. This architecture could serve as the net-zero
solution as long as the diesel gensets are maintained off the grid unless needed for an
emergency situation.
28
Figure 10. Current SNI Energy Structure.
29
Figure 11. Optimized and Net-Zero SNI Energy Structure.
Figure 12 represents a user interface for the MS Excel optimization. In this model,
black numbers are user defined, green displays MS Excel Solver calculated solutions, and
red are the constraints. Segment 1 represents the SNI energy sources in the rows and SNI
load locations in the columns. The actual distances between them are estimated from the
infrastructure map shown in Figure 5. Segment 2 represents potential capacities of the
current and proposed energy sources on SNI. Segment 3 provides the estimates of the total
cost (such as investment, operation, maintenance) required to deliver energy to the loads.
This requires calculations based on generation source, distance from the load, efficiencies,
30
losses, and more. While the exact numbers are not known, this study used a simple
approach to estimate them. This approach was based on a simple relationship (Duke, 2017):
𝜌𝐿
𝑅=
𝐴
relating the resistance R (in ohms), resistivity𝜌 is (in ohms*m), the distance
between the source and the load L (in meters), and the cross-sectional area of the
transmission wire A (in 𝑚2 ).
Figure 12. MS Excel SNI Optimization Interface.
Obviously, the greater the distance of the load from the energy generation source,
the greater the resistivity. The greater the resistivity, the greater the losses in the
transmission from source to load. The greater the losses from source to load, the more it
costs to deliver a unit of energy. Assuming 𝜌𝐴−1 is constant, the cost is therefore inversely
31
1
proportional to the distance, i.e. 𝐶 ~ . Segment 4 shows the maximum number available
𝐿
of each source in the “current,” “optimized,” and “net-zero” cells. Additional

configurations can be added in columns that follow, if needed. These cells are calculated
and constrained as integers by the MS Excel Solver. The operating hours (per day) of the
load at each location appear in “req” line. These were estimated based on a typical day of
operations at SNI. For example, Nick Town is continuously manned, so there needs to be
reliable power delivered there 24 hours per day. However, the airfield, radar, and telemetry
sites are not always in operation as often as Nick Town and thus have a lower power
requirement. Segment 4 contains the variable cells that displays the number of energy
sources required to meet a specific configuration. For this thesis, the three configurations
are shown in yellow shaded Constraints box. Segment 5 contains the required hours of
operation where energy supply is required, the actual hours obtained through the selected
simulation (to ensure the need is being met), and the percent RE of the model. The values
in Segment 2 are again used as part of the calculation for the actual and RE fraction rows
in Segment 5.
To operate the developed model, the user needs to select the MS Excel Data tab,
then the Solver button. The blue objective function, “total cost,” is minimized by changing
the green variable cells. Constraints include the red “actual” cells being greater than or
equal to the black “req” cells and the red “Solution sums” cells being less than or equal to
the black “current” “optimized,” or “net-zero” cells on a case by case basis. That is, the
model must be run three times to solve each current, optimized, and net-zero case by
changing the constraint as appropriate in Solver to match the model being run. In this case,
the model was run three times for the three different configurations and Figure 12 displays
the results from the net-zero model. More can be added as desired. Ensure that the box for
“Make Unconstrained Variables Non-Negative” is checked and that the “Solving Method
as Simplex EP” is chosen.
Running the model three times, once for each configuration (current, optimized,
and net-zero), produces the results shown at the bottom of Figure 12. The numbers of
required energy sources, total cost, and % RE are displayed. It should be noted that the
terms “current” and “reference” are used interchangeably in terms of these models and in
32
EnergyPLAN. In terms of capacities, wind turbines are 100 kW each, solar PV is 1 MW
each, and batteries are 250 kWh each.
As shown, the net-zero model yields the lowest relative cost as compared to the
current and optimized configurations. This calculation does not reflect the investment cost
to achieve a net-zero status. That will need to be considered based on the needs of the
facility and then a break-even point can be calculated. The main cost driver in the current
and optimized configurations is the use of the diesel gensets. This is the reason for the cost
increase for the optimized model as compared to the reference model. Here, both the
traditional diesel genset source and the RE sources are being used, but not to optimal
efficiency, thus the cost is increased. It is important to note that the loads on SNI can be
fully carried in all three models. The choice for which model to employ will be up to the
decision maker based on a number of factors, most importantly being the budget for
investment.
D. ENERGYPLAN REFERENCE MODEL
Next, EnergyPLAN was used to take the MS Excel optimized outputs for each
model and derive total system, annual energy balances for each configuration as shown in
Table 2. That is, the results for each model from Figure 12 were used as direct inputs into
EnergyPLAN as the number of each source of energy. The results of these configurations
are shown graphically in Chapter V. Again, all three configurations meet the current SNI
demand. Also, all five diesel gensets are not required in any of the models to meet the SNI
demand. This correlates with the manner in which the diesel gensets are currently operated
on the island, which also happens to not be the most efficient method.
Table 2. Reference, Enhanced, and Net Zero Model Configurations.

number of units | capacity
Model Diesel Wind Solar Storage
Current 2 | 1000 kW 7 | 700 kW - -
Optimized 2 | 1000 kW 7 | 700 kW 1 | 1 MW 2 | 500 kWh
Net-Zero - 9 | 900 kW 2 | 2 MW 2 | 500 kWh
33
A few assumptions for these models include:
 At any given time, power available from generated and stored energy
sources must meet or exceed total SNI demand.
 Excess power generated will be converted to stored energy via lithium ion
batteries in the optimized and net-zero configurations.
Given a mix of generators and storage capacity, the goal of the EnergyPLAN
simulation and optimization is to minimize fuel consumption by managing the mix of
generators running at a particular time and by managing the batteries charging and
discharging times.
While EnergyPLAN is fully capable of complex energy system modeling solutions,

the steps provided in Appendix A outline the basic required steps to construct a reference
model. It is the hope of the author of this thesis that the steps outlined in Appendix A will
be taken to optimize energy grids at other isolated and disparate U.S. locations and achieve
as close to a net-zero solution as possible at those sites.
If the data is more than hourly, some type of filtering will be required in order to
meet the 8784 points required by EnergyPLAN. MATLAB is the best solution for this
processing. Another option, but much more labor intensive, is using MS Excel. If
MATLAB is the method of choice, this will also afford the opportunity to be able to run
certain optimizations later in the analysis by using the MATLAB wrapper for
EnergyPLAN. This was designed by Pedro J. Cabrerra of the University of Las Palmas de
Gran Canaria and can be used to call EnergyPLAN directly from MATLAB. This is a
powerful extension of an already capable energy analysis software package and is free to
download from the EnergyPLAN website.
If annual data is unable to be obtained or there are fewer than the required 8784
data points, the data that is on hand can simply be replicated with knowing some amount
of error and uncertainty has been introduced. While this is not ideal and will lead to a less
than an optimal solution, it will provide a solid starting point to assess the energy situation
and way forward.
34
V. ANALYSIS
This chapter presents the analyses of the SNI reference, optimized, and net-zero
models as well as an energy source sensitivity analysis.
A. REFERENCE MODEL
The results of the EnergyPLAN simulation on the reference (current) model

configuration are shown in Figure 13. This model serves as the SNI reference model, to
which the optimized and net-zero models will be compared. This model represents the
island as it is today, with only diesel generators and wind turbines supplying the island
loads. The red area in Figure 13a, labeled Consumption, depicts SNI demand over a full
year. The x-axis shows hourly data points (8784) for a full year. As seen, average daily
demand (~525 kW) is relatively constant throughout the year. Note that the plot appearance
in EnergyPLAN is not editable. Hence, the other entries shown in the legend box refer to
non-existing data entries for these models.
Due to the heavy reliance on diesel gensets, this is the least efficient and most costly
of the three models under analysis in this thesis. As shown in Figure 13b, production
(shown in yellow) far exceeds demand (shown in red in Figure 13a) and much opportunity
to send energy to storage is lost. As shown in Figure 6 of Chapter III, much of this
production is simply directed to a dump load due to no storage capability at SNI. Yellow,
labeled PP+ (power plant), represents the diesel gensets and the dark blue, labeled RES12
(renewable energy source), represents wind turbine production. Data points were collected
from August to July; therefore, the wind production appears to peak midyear, but that is
actually the winter months that fall in the middle of the data collection period. These are
the windiest months at SNI and yield the most productive results for the wind turbines. The
lost opportunity is shown in Figure 13c. The goals for the subsequent models will be to
find a more optimal solution for the total SNI energy balance and attempt to achieve a net-
zero solution. As noted before, it should be noted that all of the other colors are not editable
and are built into the EnergyPLAN software. These are default EnergyPLAN assignments
and cannot be changed by the user.
35
a)
b)
c)
Figure 13. SNI Reference Model Annual Energy Balance.
36
B. OPTIMIZED MODEL
The results of the EnergyPLAN simulation on the optimized model configuration

are shown in Figure 14. From Figure 12 and Table 2 in Chapter IV, this model adds a 1MW
solar PV farm and a 500-kWh Tesla storage battery to the reference model. The SNI
demand graph, the red area in Figure 14b remains the same, at ~525 kW per day. As shown
in Figure 14b, this model depicts a lower reliance on the diesel gensets, again shown in
yellow under PP+, than the reference model due to the solar PV and the storage battery
(shown in dark blue under RES12) carrying more of the total SNI load. However, as shown
in Figure 14c, total production again exceeds demand, and there is still an opportunity to
refine the configuration.
This model is not the recommended solution for SNI energy needs, but is shown as
a step between the current SNI configuration and the possibility of a net-zero solution on
the island. Other forms of RE can be explored here as well, such as wave power,
compressed air energy storage, hydrogen fuel cells.
See Appendix A for a step-by-step guide on how to enter data in EnergyPLAN and
achieve these and the other output graphs.
37
a)
b)
c)
Figure 14. SNI Optimized Model Annual Energy Balance.
38
C. NET-ZERO MODEL
Again, from Figure 12 and Table 2 in Chapter IV, the net-zero model represents
SNI with 2 MW of solar PV, 900 kW of wind turbines (a 200 kW increase over the
currently installed wind capacity), and a 500-kWh Tesla storage battery. Nine 100 kW
wind turbines, 2MW of solar PV, 500 kWh Tesla storage battery, and zero diesel gensets
is the full solution to a net-zero SNI. The average daily SNI demand (~525 kW) is again
shown in red in Figure 15a for analysis. As shown, this model is fully capable of meeting
the average daily demand with excess to account for any surge demand times. Figure 15b
shows the RE generation exceeding the demand in Figure 15a at many times throughout
the year. Where there is excess production, this is sent to battery storage and subsequently
used when the production falls below demand. The yellow PP+ portion of the plot is
maintained for comparison purposes only.
Figure 15c shows the SNI total energy balance achieving a net-zero solution,
several times throughout the year. If the storage battery was preloaded with a decent charge,
this solution would have been achieved even sooner. This SNI model is also able to increase
the stabilization of the grid due to the large percentage of RE share. The exact numbers of
installed components can be increased or decreased with the MS Excel optimization tool
in order to allow for more or less overshoot to properly balance the grid and the storage
battery. Or, perhaps, the excess can somehow be exported as a source of revenue.
39
a)
b)
c)
Figure 15. SNI Net-Zero Model Annual Energy Balance.
40
D. SENSITIVITY ANALYSIS
This section completes the discussion with a sensitivity analysis of the three
proposed energy models. Since some months afford more of an opportunity to harvest more
RE production than others, it is appropriate to consider the effects of not fully employing
all of the proposed configurations from the MS Excel optimization. The winter months
tend to be the windiest at SNI and will therefore be more reliable for wind turbine
employment. As for solar PV, and not considering cloud cover, the island is nearly
completely treeless and should be an optimal spot for year-round solar irradiance.
As seen from Figure 16a, representing the reference model, the diesel gensets
comprise the majority of the energy production capacity. While SNI has five diesel gensets
(Table 1), only two of them are in operation at any given time. Therefore, losing one diesel
genset would account for a 30% drop in SNI power supply. This would be a drastic
situation for the island and the energy needs would be severely impacted. For comparison,
losing a single wind turbine accounts for only a 5.9% drop in supply. Hence, not all of the
reference model is flawed, it is just not robust enough to account for a loss of a diesel
genset.
As seen from Figure 16b, for the optimized model, the diesel gensets and the wind
turbines are nearly identical in share of supply at 42% and 42.7% respectively. Here, losing
a diesel geneset is even more drastic due to the MS Excel optimization only allowing one
diesel genset to operate on the grid at a time. Losing a single wind turbine though is still
only a 5.9% drop in supply. Solar PV carries 12.2% and the storage battery shows a 3.1%
share. The diesel genset is still vital to SNI energy needs in this model.
Finally, for the net-zero model, shown in Figure 16c, the diesel gensets are
completely removed from the grid and the wind turbines are now carrying 73.7% of the
load. Since there are now nine wind turbines, losing a single turbine accounts for an 8.2%
drop in supply. This could be significant in the summer months which typically feature
weaker winds. Solar PV carries 10.5% and the storage battery shows a 2.6% share.
Additional solar PV or battery storage can be considered to lessen the potential impact of
a loss of multiple wind turbines in the summer months.
41
Reference Model
80%
Power Supply %
60%
40%
Wind Turbines
20%
Diesel Gensets
0%
0 2 4 6 8
Number of Sources
a)
Optimized Model
50%
Power Supply %
40%
30% Wind Turbines
20% Diesel Gensets
10%
Solar PV
0%
0 2 4 6 8 Battery
Number of Sources
b)
Net-Zero Model
80.0%
Power Supply %
60.0%
40.0% Series1
20.0% Solar PV
0.0% Battery
0 2 4 6 8 10
Number of Sources
c)
Figure 16. Sensitivity Analysis for Reference, Optimized, and Net-Zero Energy
Models
42
VI. CONCLUSIONS AND RECOMMENDATIONS
A. CONCLUSIONS
This thesis resulted in the following elements:
 a fully modeled analysis of installed energy capacity, demand, balancing,

and cost at SNI,
 a step-by-step guide in the form of the FIDE-USN that can be used to

replicate this analysis at similar U.S. Navy locations,
 varying technical simulations and analysis of the SNI total energy balance,
 technical multifactor optimizations of the generation and storage models,
 microgrid architecture design analysis and proposals, and
 systems level thinking in terms of the strategic (mission assurance),

technical (efficiencies of microgrid), and financial considerations.
It was shown that the location, operations, and energy requirements of SNI provide
a unique opportunity for the DOD to achieve a tremendous benefit as a first of its kind net-
zero facility. The current day-to-day costs associated with maintaining SNI are measurably
higher than similarly sized facilities in the continental United States. All fuel and food are
transported to the island, and waste is transported off the island. This supply chain also
results in an island community that is very vulnerable to supply interruptions and is
completely dependent on outside supplies for survival. The baseline use of resources and
strategies for reducing energy use and providing energy from on-site renewable energy
systems have been described. Findings indicate that not only could on-site systems provide
much of the island requirements, but also that it may be cost effective to do so. It would be
possible to make the island self-sufficient, or nearly so, in terms of energy use.
The opportunity at SNI is both unique and significant. The realization of energy
security and resilience is enormous for the U.S. Navy and SNI provides that opportunity.
43
One of the main goals of this thesis was to show and demonstrate that a military installation
can be 100% RE. From there, it is possible to scale and adjust the solution to other sites, as
appropriate. If a 100% RE solution is achieved at SNI, it would be the first know instance
in the DOD. As mentioned in Anderson and Yakimenko (2017, 2), other notable global
locations with high levels of RE include El Hierro in the Canary Islands (Spain) at 100%
penetration with 10,000 people on the island. Isle of Eigg (Scotland) at 94% penetration
with 94 total people on the island. And, Samsø (Denmark) at 100% penetration with an
estimated 3,000 to 4,000 inhabitants.
The modeling and analysis efforts allowed the four key research questions to be
addressed as presented in Table 3.
Table 3. Research Questions and Conclusions

Research Question Conclusion
Is it possible to accurately model and apply Yes, this was proven to be possible for SNI
power system microgrid technology at with a step-by-step guide created for use at
U.S. Navy remote and disparate island other locations.
facilities?
Can the derived model lead to a more Yes, the proposed models can lead to a
efficient total system energy balance for more efficient and potentially net-zero
San Nicolas Island and other U.S. Navy U.S. Navy installation at SNI. This would
islands? be a first for the DOD.
Can multifactor optimizations enhance the Yes, but more work is needed to further
proposed models and lead to enhanced refine the optimizations.
control solutions for the power system
microgrids?
Can U.S. Navy bases become more energy Yes, if a facility is able to achieve a net-
self-sufficient to the point of increasing zero situation.
energy security and resiliency?
B. RECOMMENDATIONS
The goal for future work is to optimize microgrids for remote islands even further
by using complex systems tools. EnergyPLAN, MATLAB, Solver and Agent Based
44
Modelling (either through NetLogo or MATLAB) and applying these tools to isolated U.S.
Navy locations could be a discrete tool and valuable to the DOD. Through modeling
facilities with these tools, synthesizing the results, and then revising the control system
solution in MATLAB, a greater efficiency can be gained approaching a 100% RE solution.
The following areas offer opportunities for further research in the important field
of U.S. Navy energy microgrids:
 Apply the analysis presented in this thesis to the following U.S. Navy
locations: Diego Garcia, Guantanamo Bay, San Clemente, Camp
Lemonnier (Djibouti).
 Monitor SNI electricity demand and supply in real-time, similar to Canary

Islands (Real Time Demand and Generation, 2017).
 Use agent-based modeling (ABM) through MATLAB or NetLogo.
 Employ control systems dynamics modeling.
 Apply non-linear energy system modeling.
 Construct a predictive control model based on weather.
45
46
APPENDIX A. FINDING AND INPUTTING DATA INTO
ENERGYPLAN—UNITED STATES NAVY (FIDE-USN)
The goal of the FIDE-USN is to assist others working on green energy solutions for
U.S. Navy installations on the processes required to maximize the EnergyPLAN software
model. I hope that my lessons learned in modeling SNI will help others more efficiently
find the correct data and enter it efficiently into EnergyPLAN. This appendix will assist in
the creation of a reference model for any U.S. Navy facility’s energy system from which it
can then be compared to alternatives by adding in renewable green energy solutions to the
model. While EnergyPLAN is fully capable of complex energy system modeling solutions,
the steps below outline the basic required steps to construct a reference model.
1. Sign up for a free account at http://www.energyplan.eu/.
2. Download and install the EnergyPLAN software. This only works on a PC. If
you are working on a Mac, you will need to download and install some form of virtual
machine. I recommend VMware Fusion.
3. Acquire the following technical data regarding the specific site or facility:
Note: The following data collection steps may require local data collection or, if
available, downloading data from a remote networked computer. You will need 8784 data
points to build the reference diesel genset data file for EnergyPLAN. This, as with wind
and solar data, is equivalent to hourly data for a year. If you have less data than this, you
may need to simply replicate the data to meet the 8784 data point requirement. Work with
what data is available but strive for 8784 unique data points in order to construct the most
accurate model.
 diesel genset hourly data for a year
 wind turbine hourly data for a year
 solar photovoltaic hourly data for a year
4. The total annual production/demand (TWh/year).
47
5. Find the installed supply capacities (MW) of all supply sources at the site.
6. Find or best estimate the following costs at the site in $USD:
 Overall machine and material investments
 Fixed operations and maintenance
 Variable operation and maintenance
 Fuel(s)
 Transportation (vehicles)
7. Collect a year of weather data. To get a full report:
 I recommend using https://www.wunderground.com/ and search for the

target location
 Go to the historical data as follows: History  Custom  enter dates 

Get History
 Export this data to MS Excel. This may require select all, copy, and paste.
8. Enter all of the above data into EnergyPLAN as follows:
 Total annual production/demand: Demand tab  Electricity demand field.
 Diesel Genset distribution: Enter the 8784 data point distribution in the
line next to the Electricity demand box by selecting the “Change
distribution” button.
 Wind 8784 data point distribution: Supply  Electricity Only 

Intermittent Renewable Electricity  select the appropriate dropdown. Be
sure to enter the Capacity kW and enter the distribution by selecting the
“Change” button to the right.
 Follow the same process for Solar PV as for Wind.
48
9. Run the reference model by going to the Output Tab and then selecting Graphics.
Use the embedded page buttons to achieve your desired graphical result.
10. Select the Run (Screen) button from the navigation bar to acquire the numerical
report.
11. Start adding in proposed RE energy sources and analyzing how your model
changes in terms of both technical outputs and economic outputs. If unable to accurately
determine the correct number of RE sources to add to the model, consider using the MS
Excel Optimization techniques presented in Chapter IV.
Notes:
1. It is vital to save all distribution files as Windows Formatted Text (.txt). This is
the file type required for EnergyPLAN. The distribution is inputted as a text file and stored
in the “Distributions” folder of EnergyPLAN.
2. Specifically for the annual distributions files that contain 8784 data points,
Connolly (2017, 5) states in the FIDE Guide:
The data points are usually between 0 and 1, representing 0–100% of

production or demand. However, if a distribution is entered with values
greater than 1, EnergyPLAN will automatically index the distribution. This
is done by dividing each entry in the distribution by the maximum value in
the distribution. One exception is the price distribution under the
‘Regulation’ tab, which does not index the inputs.
3. Prior to adding battery storage capacity, make sure to understand how

EnergyPLAN treats storage batteries. Connolly (2017, 38) states in the FIDE Guide:
In EnergyPLAN, electricity storage is described in the form of pumped

hydroelectric energy storage (PHES) as this is the largest and most common
form of electricity storage in use today. However, this can be used to define
any type of electricity storage which has a charging capacity
(pump/compressor), discharge capacity (turbine), and a storage capacity.
When defining the electricity storage capacities available, it is also possible
to define an electricity storage operation strategy.
49
50
APPENDIX B. DATA PREPROCESSING MATLAB CODE
Figure 17. MATLAB Code Used to Process SNI Energy Demand Data Files.
Source: Yakimenko (2017).
51
52
LIST OF REFERENCES
Anderson, Bill, Michael Cruz, Nicholas DeMarco, Bruce Garrett, and Nathanael Royer.
2017. A Microgrid Analysis of San Nicolas Island Using HOMER. Port Hueneme,
CA: Naval Facilities Engineering Command (NAVFAC) Engineering and
Expeditionary Warfare (EXWC).
Anderson, William W., and Oleg A. Yakimenko. “Comparative Analysis of Two

Microgrid Solutions for Island Green Energy Supply Sustainability.” IEEE
International Conference on Renewable Energy Research and Applications. San
Diego, CA, 2017.
Connolly, David. 2015. Finding and Inputting Data into EnergyPLAN (the FIDE Guide).
5th ed. Denmark: Aalborg University.
Duke University Department of Physics. 2017. “Resistance and Resistivity.” Accessed

November 1, 2017.
http://webhome.phy.duke.edu/~kolena/resistance_resistivity.pdf
EnergyPLAN, n.d. “Advanced Energy System Analysis Computer Model.” Accessed

July 5, 2017. http://www.energyplan.eu/.
Google Maps. 2017. “San Nicolas Island.” Accessed August 1, 2017.

https://www.google.com/maps.
Harley, William. 2016. “Redefining Security: System of Systems Survivability

Assessment Model for Life Cycle Security and Attack Vector Design.” Master’s
thesis, Naval Postgraduate School.
International Council on Systems Engineering. 2017. “Working Groups.” Accessed July

18, 2017.
https://www.incose.org/ChaptersGroups/WorkingGroups/Application/power-
energy-systems.
Kandt, Alicen, and Andy Walker. 2008. San Nicolas Island, CA: Renewable Community
Plan Outline, Baseline Development, and Initial Renewable Assessment. Golden,
CO: National Renewable Energy Laboratory.
Lund, Henrik. 2015. EnergyPLAN: Advanced Energy Systems Analysis Computer Model.
12th ed. Aalborg University, Denmark. Sustainable Energy Planning Research
Group.
Lund, Henrik, and David Connolly. 2016. EnergyPLAN: Energy System Analysis Tool.
Vol. 12.5. Aalborg University, Denmark. Sustainable Energy Planning Research
Group.
53
National Renewable Energy Laboratory. 2017. “NSRDB Data Viewer.” Accessed August
3, 2017, https://maps.nrel.gov/nsrdb-
viewer/?aL=8VWYIh%255Bv%255D%3Dt%268VWYIh%255Bd%255D%3D1
&bL=groad&cE=0&lR=0&mC=31.98944183792288%2C-105.99609375&zL=4
Olsen, Timothy L., and Ed McKenna. 1996. Hybrid Energy System Cost Analysis: San
Nicolas Island, California. Golden, CO: National Renewable Energy Laboratory.
Red Eléctrica de España. 2017. “Real-Time Demand and Generation.” Accessed July 28,
2017, http://www.ree.es/en/activities/realtime-demand-and-generation.
Roth, John, P. 2017. “Fiscal Year (FY) 2018 Fuel Price Change.” Official Memorandum.
Washington, DC: Department of Defense.
Science Alert. 2017. “This Island in American Samoa Is Almost 100% Powered by Tesla
Solar Panels.” Accessed October 27, 2017, https://www.sciencealert.com/this-
island-in-american-samoa-is-almost-100-powered-by-tesla-solar-panels.
Scott, Zane. 2016. 9 Laws of Effective Systems Engineering. Blacksburg, Virginia: Vitech
Corporation.
Sugiyama, Yutaka. 2017. Modular Micro Grid (M2G). Office of Naval Research Energy
Systems Technology Evaluation Program FY17 Review, Monterey, CA, July 25,
2017.
Today in Energy. 2017. “U. S. Energy Information Administration.” Accessed July 18,
2017. https://www.eia.gov/todayinenergy/archive.php.
Weather Underground. 2017. “San Nicolas Island, CA.” Accessed July 19,
2017, https://www.wunderground.com/history/airport/KNSI/2017/1/1/CustomHist
ory.html?dayend=19&monthend=7&yearend=2017&req_city=&req_state=&req_
statename=&reqdb.zip=&reqdb.magic=&reqdb.wmo=.
Yakimenko, Oleg. 2017. “MATLAB tutorial.” Class notes for Introduction to Scientific
Programming, Naval Postgraduate School, Monterey, CA.
54
INITIAL DISTRIBUTION LIST
1. Defense Technical Information Center

Ft. Belvoir, Virginia
2. Dudley Knox Library

Naval Postgraduate School
Monterey, California
55

Naval Postgraduate School: Monterey, California

Uploaded by

Copyright:

Available Formats

Naval Postgraduate School: Monterey, California

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Naval Postgraduate School: Monterey, California

Uploaded by

Copyright:

Available Formats

NAVAL

MODELING AND OPTIMIZING GREEN MICROGRIDS

Thesis Advisor: Oleg A. Yakimenko

Approved for public release. Distribution is unlimited.

14. SUBJECT TERMS 15. NUMBER OF

MODELING AND OPTIMIZING GREEN MICROGRIDS AT REMOTE U.S.

Submitted in partial fulfillment of the

MASTER OF SCIENCE IN SYSTEMS ENGINEERING

NAVAL POSTGRADUATE SCHOOL

Approved by: Oleg A. Yakimenko

II. MICROGRIDS AND TOOLS FOR THEIR ANALYSIS ..................................7

III. SAN NICOLAS ISLAND ENERGY SYSTEM .................................................15

IV. DATA COLLECTION AND MODELING .......................................................23

VI. CONCLUSIONS AND RECOMMENDATIONS .............................................43

APPENDIX A. FINDING AND INPUTTING DATA INTO ENERGYPLAN—

APPENDIX B. DATA PREPROCESSING MATLAB CODE ....................................51

LIST OF REFERENCES ................................................................................................53

INITIAL DISTRIBUTION LIST ...................................................................................55

Figure 1. Microgrid Components. Source: Anderson and Yakimenko (2017). ..........7

Figure 2. Typical Energy System in EnergyPLAN. Source: EnergyPLAN

Figure 3. California Channel Islands. Source: Google Maps (2017). .......................17

Figure 5. Official U.S. Navy SNI Infrastructure Map. Source: NAVFAC

Figure 8. Direct Normal Irradiance Graph. Source: NSRDB Data Viewer

Figure 10. Current SNI Energy Structure....................................................................29

Figure 11. Optimized and Net-Zero SNI Energy Structure.........................................30

Figure 12. MS Excel SNI Optimization Interface. ......................................................31

Figure 13. SNI Reference Model Annual Energy Balance. ........................................36

Figure 14. SNI Optimized Model Annual Energy Balance.........................................38

Figure 15. SNI Net-Zero Model Annual Energy Balance. ..........................................40

Table 1. SNI Diesel Gensets. ...................................................................................16

Table 2. Reference, Enhanced, and Net Zero Model Configurations. .....................33

Table 3. Research Questions and Conclusions ........................................................44

CAES compressed air energy storage

DNI direct normal irradiance

DON Department of the Navy

DOD Department of Defense

ESTEP Energy Systems Technology Evaluation Program

FIDE Finding and Inputting Data into EnergyPLAN

FIDE-USN Finding and Inputting Data into EnergyPLAN – U.S. Navy

HOMER Hybrid Optimization Model for Multiple Energy Resources

INCOSE International Council on Systems Engineering

NPS Naval Postgraduate School

NAVFAC EXWC Naval Facilities Engineering and Expeditionary Warfare Center

NREL National Renewable Energy Laboratory

ONR Office of Naval Research

PESWG Power and Energy Working Group

PSM Physical Solar Model

RSWG Resilient Systems Working Group

SNI San Nicolas Island

SoS system of systems

This thesis produced the following:

 A fully modeled analysis of energy capacity, demand, balancing, and cost

 A step-by-step guide in the form of the FIDE-USN to replicate the

 Various technical simulations and analysis of the SNI total energy

 Systems-level thinking in terms of the strategic (mission assurance),

Figure 1. SNI Net-Zero Model Results.

Anderson, William W., and Oleg A. Yakimenko. “Comparative Analysis of Two

A. BACKGROUND AND PROBLEM STATEMENT

The International Council on Systems Engineering (INCOSE) believes that power,