This section provides background on several topics that are needed to understand the contribution of this paper. Further, related research is discussed and gaps are identified which this paper aims to fill.
2.1. Military Microgrids
Many military bases and other critical facilities have or are in the process of implementing microgrids, at least in part in an attempt to provide reliable energy to critical loads. Microgrids on many military bases generally perform six key functions for the attached critical loads including: (1) stepping down high voltage power received from the main utility grid to intermediate and/or low voltage, (2) distribute energy to critical and non-critical loads, and connect other electrical hardware, (3) generate energy locally (e.g., diesel gensets, Photovoltaic (PV) arrays, wind turbines, etc.), (4) store energy (e.g., battery banks, pump-storage hydro, etc.), (5) control the flow of energy, the generation, the storage, and the loads throughout the microgrid using one or more controllers to automatically operate electrical equipment, (6) step up and down voltage (transformers) and convert energy from AC to DC or vice versa (inverters and converters) [
12,
13]. The main components of a microgrid often include diesel generators and other fossil fuel generators; PV, wind turbines, and other renewable energy sources; energy storage systems such as battery banks; Points of Common Coupling (PCC) where the microgrid connects to the utility grid (generally with a switch to allow for operation of the microgrid while disconnected from the utility); one or more control systems (often involving Supervisory Control and Data Acquisition (SCADA) systems); and switches, converters, inverters, relays, transformers, power lines (above ground, underground, etc.), and other related hardware.
An established PCC and other requisite hardware (grid-forming generation and/or storage, generation, local controller(s), etc.) provides one of the primary benefits of microgrids for military bases and other critical facilities. The PCC provides the ability to disconnect from the main utility grid, called “island-mode” [
14], and continue uninterrupted service of critical loads. In island-mode, loads critical to a military base’s primary mission functions (the critical loads) are still provided power from local generation and storage sources [
14]. Issuances and instructions from organizations within militaries such as the NAVFAC generally state that in order for military bases to perform their missions, energy security of critical loads must be ensured regardless of the state of the grid beyond the PCC and outside of the microgrid system boundary [
8]. One way of identifying critical loads in the United States military is the Mission Assurance and Continuity of Operations plan for DOD installations which instructs that the essential buildings needed to conduct national security missions based on criteria identified by the individual branches of the military (e.g, DoN, etc.) and the installation mission must continue to receive power even during utility disruptions (e.g., external power to a base is cut) [
8].
Military microgrids have found several benefits in use including integrating smart grid technologies, reducing peak load and losses by enhancing integration of Distributed Energy Resources (DERs) (often including diesel gensets, wind turbines, micro hydro generation, fuel cells, PV, and other generation sources), localizing power quality and reliability for end-user satisfaction, and supporting the utility grid by managing sensitive loads and variability of DERs [
10,
15]. Currently military microgrids are allowing the interconnection of loads and DER that can replace duplicitive and expensive-to-maintain small Uninterrupted Power Supply (UPS) and stand-alone backup generators [
10]. The flexible architecture of microgrids eases the employment of DER in conjunction with controllable loads and storage devices [
15,
16].
2.2. Nanogrids
Nanogrids are generally much smaller than microgrids. While an average military microgrid may serve several dozens or hundreds of loads and operate in the 5–500 megawatt range, nanogrids often serve on the order of one to five loads and a few kilowatts up to 5–10 megawatts. Nanogrids generally are technologically simpler than microgrids because they only serve a single building or a few loads [
17]. The smaller, simpler design of nanogrids usually allows power production to occur much closer to the point of consumption versus grid infrastructure and even military microgrids which sometimes can encompass many hundreds of square kilometers of service territory. Generation occurring at the point of use significantly reduces the potential negative effects of transmission and distribution lines such as loss of efficiency, single point of failure, etc. [
18]. Certain nanogrid configurations (discussed later) where the nanogrid is connected to a microgrid or grid via its own PCC can be very fault-tolerant where they are able to successfully transition to island-mode and continue serving critical loads until microgrid and/or grid power is restored. Many of the same benefits found with the implementation of military microgrids are expected to be realized with the deployment of nanogrids on military bases.
It is important to note common technical characteristics of nanogrids. Nanogrid voltage levels are often lower than microgrids and often are in the 100–500 volt range for both AC and DC nanogrids. Power generation and consumption is often much smaller in a nanogrid than a microgrid with ranges from 1 watt for the smallest nanogrids powering extremely small loads up to 1 megawatt; however, the upper limit can vary and is defined by the entity implementing the nanogrid [
19]. Typical nanogrid loads are on the scale of a single appliance or computer up to a building [
17]. Nanogrids can sometimes be used to improve electrical efficiency by 5–13% in some residential applications versus other options [
19]. A common nanogrid system architecture uses DC because using DC can be more efficient due to many nanogrid power sources producing DC power and many storage systems (Energy Storage System (ESS)) using DC power which results in fewer inverters and converters [
18,
19]. However, while DC nanogrids can increase efficiency, they also come with the consequence of needing enhanced protection against short circuit line faults and ground faults [
18]. The mitigation of faults is done with arching-type circuit breakers or more advance tactics that involve special mechanical circuit breakers that open against fault currents by forcing currents to zero by external means to extinguish arcs [
20].
2.3. Energy Resilience
The focus of this paper is on improving energy resilience of military bases, and specifically for critical loads, through the implementation of nanogrids. In order to understand the amount of improvement in resilience a proposed nanogrid may have over existing microgrid infrastructure, a working definition of resilience in this context and a quantitative means of measuring resilience must be defined.
Rather than adopting a civilian-focused energy resilience definition and quantification, it is important to first understand the value of and difference in resilience from a military perspective. Civilian grid systems often focus on defining resilience in terms of real dollars lost when energy supply does not meet demand and a facility stops production of something that is easily monetized (e.g., steel, automobiles, computer chips, processed food, vaccines, etc.) [
21]. Conversely, military energy systems such as nanogrids and microgrids produce something less tangible: national security [
22]. National security is intangible and has no easily defined value [
22,
23]. The cost to national security due to lack of energy resilience becomes subjective and theoretical [
24].
There are a variety of definitions of energy resilience within military communities although they all focus on several commonalities: preparing for an event, riding out the event, stabilizing after the event, and recovering from the event in order to continue to support mission essential operations and maintain readiness [
5,
25]. The military definitions of energy resilience generally align with civilian definitions although the military definitions always tie back to the mission of national security [
5,
26]. Thus, this paper adopts the definition of energy resilience from a military perspective as encompassing the ability of an energy system to support critical loads before an event, during an event, immediately after an event, and in the recovery from an event back to a normal operating state.
There have been several attempts to quantify energy resilience for military purposes from a financial perspective. For instance, a cost benefit analysis of stand-alone diesel generators attached to critical loads on installations was performed to calculate a Customer Damage Function (CDF) which is representative of the cost of interruption as a function of the duration of an outage [
21,
27]. However, in most situations it is very difficult to quantify national security in a dollar amount. Instead, within the United States military, the value of resilience is sometimes defined using the Mission Dependency Index (MDI) where MDI captures the relative criticality of various infrastructure on a base with respect to the mission of the tenant organizations on a base on a 0–100 scale with 100 being absolutely critical [
28]. In contrast, some researchers have criticized the use of MDI when directly ranking criticality of loads to their overall role in national security missions. These researchers claim existence of inaccuracies in addressing time dependency of corrective actions, and misrepresentation of mission interdependence and intradependence in the MDI equation [
29]. Recently the DoN has begun using Resilient Energy Program Office (REPO) (an attempt to address the shortcomings of MDI with similar objectives and using aspects of Energy Security Assessment Tool (ESAT)) as a replacement in some of the roles that MDI has been previously employed [
30]. However, many organizations across the United States federal government retain MDI and some issues with REPO are currently being identified. Thus, this paper adopts MDI as the base measure of a unit of energy resilience.
In order to use MDI to quantify energy resilience for the military, a method of quantifying resilience over time is needed. Several exist or are in development [
5,
22,
26,
31,
32]. This paper adopts the approach proposed by Peterson et al. [
22] where the Expected Electrical Distribution Mission Impact (EEDMI) quantifies the resilience of an energy system versus all expected initiating events, threats, disruptions, etc. In Peterson et al.’s approach, MDI is used to understand the value of each critical load to national security where it provides the input to Mission Impact (MI) on a per unit time (T) basis. MI is the impact to a mission on a per unit time basis for if a specific electrical load is not served. A single scenario (specific initiating event, threat, etc.),
, is defined by the MI per unit time (T) that is
not served electricity throughout the duration of the scenario [
22],
During normal operations and with no electrical interruptions, is zero. During a scenario where MDI = 50, the unit of time is hours, and power is not delivered for 2 h, . In scenarios where not enough power is available to serve all loads, load shedding occurs. The details of how load shedding occurs (e.g., which loads are shed first, rotating blackouts, etc.) is dependent upon behavior of the energy system and controller(s). In such situations where load shedding occurs, which indicates an impact to national security and thus warrants further investigation.
The aggregate of all scenarios (
S) calculated in Equation (
1) is EEDMI which includes the probability (
) of the specific threat or initiating event occurring over the course of a year:
The total EEDMI value for a specific electrical system configuration is used as a way to compare between different potential electrical system architectures from the perspective of energy resilience. A lower EEDMI value is more desirable because it means that the electrical system is less susceptible to initiating events and threats disrupting power to critical loads [
26]. The process of developing EEDMI is the same whether analyzing a very small nanogrid or a very large microgrid on a military base.
2.5. Nanogrid System Design
A variety of nanogrid system designs are proposed in the literature and have seen limited implementation in the real world. Different nanogrid designs can serve different purposes and will have different impacts on the energy resilience of an electrical system. From a systems engineering perspective, Giachetti et al. advances six criteria for understanding electrical systems used to power military loads: (1) System Purpose (2) Stakeholders (3) System Boundaries (4) Functional Requirements (5) System Architecture (6) Operating Modes [
2].
In this paper, the nanogrid system’s purpose is to improve resilience, transmission efficiency, and ease of integration of renewable resource and energy storage. Though nanogrid resilience has not been validated, our analyses (detailed in subsequent sections) indicate that nanogrids do improve resilience over baseline microgrid infrastructure in many situations.
Military nanogrid stakeholders include base commanders, tenant commands, local energy companies, microgrid providers and contractors, and maintenance and funding organizations. Higher authorities that base commands report to (e.g., The Pentagon) are also impacted by nanogrids.
We view the physical and functional boundaries of a microgrid from a holistic perspective. The system includes physical equipment, processes, software, and people who sustaining operational effectiveness (e.g., maintenance, operations, supply chain, etc.). Nanogrid system boundaries differ from microgrid system boundaries in that nanogrids are much smaller than microgrids. However, individual nanogrids can be part of a larger microgrid from the perspective of a Systems of Systems (SoS). By considering several nanogrids as part of a SoS, individual nanogrids can be placed on various critical loads within a microgrid to work together to improve the microgrid’s energy resilience. Nanogrids generally have external inputs from maintenance organizations, fuel providers, operator organization, a microgrid, and the external grid.
The primary functional requirements from a systems engineering perspective of nanogrids are to generate, distribute, control distribution, and store energy. Where lower level functions under generate energy include generate electrical energy and adjust energy production. Distribute energy includes transmit energy, control energy flow, and convert energy. Controlling a nanogrid includes measure nanogrid state, process measurements for control decisions, and send control signals. Energy storage provides the benefit of stability of the nanogrid system when energy demand exceeds energy generation capacity. This includes store energy, release energy, and adjust energy flow.
System Architecture for a nanogrid differs from a microgrid due to its complexity and potential configurations. We suggest specific system architecture details of any individual nanogrid be determined after an analysis of alternative existing nanogrid architectures is conducted in order to identify which type(s) of nanogrid architecture(s) best benefit the energy resilience of any specific military installation.
Nanogrid operating modes include four primary modes of operation: microgrid-connected, transition-to-island, island, and re-connection. Microgrid-connected mode establishes normal parallel operations with the microgrid and (assuming a microgrid-to-grid PCC) utility grid while all distributed resources in the grid operate within IEEE standards and information is exchanged with the nanogrid controller. Transition-to-island mode represents the nanogrid’s transient state of transition between being connected to the microgrid and being fully islanded. The nanogrid must have sufficient energy storage available in addition to the ability to stabilize voltage and frequency (in the case of an AC or AC/DC nanogrid) for successful transition. A major concern of this mode is the dampening of transients in the nanogrid to avoid tripping protective devices [
2]. Island mode is when the nanogrid is operating independent of any outside energy sources, and loads are solely supported by DER and ESS with the responsibility to maintain set frequency and voltage parameters [
2]. Re-connection mode is the transition period where the nanogrid is reconnected to the microgrid. Before synchronization can occur, the frequency, voltage, and phase angle between the two must be within acceptable parameters in order for the nanogrid and microgrid to resume unified operations [
2].
2.6. Nanogrid Architectural Configurations from a Resilience Perspective
A number of major architectural configurations have been proposed in the literature and some have been implemented on a limited basis. This section discusses some of the considerations of nanogrid architectural configurations from a resilience perspective and from the perspective of other important military requirements such as efficiency and renewable energy.
The conflict is that many authors design nanogrid infrastructure differently: with centralized and decentralized control systems, the sole use of DC power, or a hybrid of DC-AC\AC-DC power conversions. Currently no actionable guidance exits on potential nanogrid configurations that may improve resilience of critical loads to outages on military installations. However, commercial solutions that stem from the demand for electrical power for space applications have led to similar refinement of existing technologies for nanogrid-like solutions. Similar to nanogrids, these new technologies address power conversion from PV arrays with management, regulation, and monitoring of electrical demand. Though these power systems are not specifically called nanogrids, their basic elements are similar: energy storage, power conversion, power management and distribution, and use by spacecraft systems [
34]. Further, the details of critical loads vary from installation to installation; thus, there is no one size fits all solution.
Research has investigated the benefits and drawbacks of implementing both AC and DC nanogrids which highlights the difference in cost between DC nanogrids (high up-front costs) and AC nanogrids (low up-front costs) [
18]. Though we explicitly do not consider cost within our research, it is important to note the added cost to augment existing military microgrid infrastructure with DC nanogrids. Some research has recently been focused on understanding the cost of increasing energy resilience from a military perspective [
26,
35]. While some military energy organizations are primarily driven by cost, we expect that cost will soon be balanced with energy resilience to better align with the high-level desire to assure that important national security missions can continue in spite of disruptions to grid and microgrid infrastructure.
Safety is another concern with military requirements and therefore protection concerns arise when choosing DC nanogrid architecture. Short circuit line and ground faults are more common at output terminals for DC nanogrid architecture then Alternating Current (AC) nanogrid [
18]. Examples of this are seen in Okinawa, Japan with experimentation done by researchers to stabilize DC power on nanogrids with three sets of bidirectional DC to DC converters (used in current and voltage regulated mode) to maintain a constant bus voltage [
17]. In addition, mitigation of these potential failure modes can occur with the use of arcing-type circuit breakers or more advanced strategies [
18]. Control strategies and control system design can have a large impact on a variety of important nanogrid requirements. Nanogrid architectures using either centralized or decentralized control provide different solutions to optimize power production and consumption to better match a load’s supply and load curves, and reduce the negative effects of intermittency [
18]. Centralized control at the microgrid level (controlling the microgrid plus any constituent nanogrids) enables a cohesive control strategy of system dynamics but provides a potential single point of failure. Though cohesion is important, militaries are generally more concerned about reliability and resilience. In addition, when a microgrid is under stress, the purpose of the nanogrid is to independently disconnect from the microgrid at its PCC and operate as an independent system in island mode. Therefore, centralized control strategies are undesirable for most military applications and decentralized nanogrid control systems that can react to threats to the uninterrupted delivery of power to critical loads is important for energy security and improves resilience. However, decentralized control of nanogrids can inhibit overall microgrid reliability due to there being many more potential independent failures of distributed nanogrid controllers over time [
18,
36].
A benefit of a DC nanogrid is the commonality of PV array and ESS output power generally being DC power [
18]. This commonality of a DC based nanogrid would allow for a smoother and efficient transition of power amongst renewable energy sources to battery storage. Though an AC based nanogrid will save money on initial cost upfront with no necessary retrofitting, added efficiency benefits of a DC based nanogrid will outweigh initial capital required criteria and benefit the military’s mission-focused requirements.
2.7. Related Research
Existing research into and deployment of nanogrid technologies to date have generally not directly focused on the ability of nanogrids to support critical loads from the perspective of energy resilience and especially for military base applications. As far as we are aware, no one has proposed using nanogrids from the perspective of improving resilience of critical loads that support national security. The majority of current nanogrid research focuses on conceptual nanogrid design and defining nanogrid infrastructure. Only a few publications have reported on nanogrids successfully implemented in real-world conditions. For instance, a nanogrid was implemented in a housing community in Okinawa, Japan where it was found that a decentralized DC-DC solution was beneficial [
17]. However, the authors noted that additional research is needed to explore limitations of decentralization and the development of higher level intelligent exchange strategies that enhance efficiency [
17].
Enhancing efficiency of nanogrids (an important aspect of energy security for militaries) occupies a significant portion of the existing literature where identified challenges include protocols, demand-side management, security, and the self-control of the overall system [
37]. One promising area of research is fully DC-based nanogrids where higher efficiency, better power quality, and better stability is achieved versus other options and may contribute significantly to energy security as a result [
18].
As mentioned previously, a similar concept to DC nanogrids is zonal distribution (sometimes referred to as zonal shipboard power) which is implemented on a growing number of vessels in many Surface Navies. Zonal distribution uses DC power architecture in order to avoid some of the issues associated with AC power architecture such as generator sets working at fixed speed that limit fuel efficiency, reactive power flow and power quality problems, bulky conventional transformers, and challenges associated with supporting pulsed electric loads which will become increasingly common on Navy ships [
38].
In addition, current research in nanogrids fails to support a mission-focused objective for DOD shore installations and instead supports a commercial cost effective mission with an ease of integration to existing infrastructure. Current research suggested the pursuit of future work in resilience of smart load configurations, networks and connections, and fast response to disasters [
18,
39]. This segues into the uniqueness of our research where we propose a methodological approach to enhance resilience of mission critical loads to create greater energy security that in turn supports the mission of national security.