The Emerging Potential of Microgrids in the Transition to 100% Renewable Energy Systems

: International, national, and subnational laws and policies call for rapidly decarbonizing energy systems around the globe. This effort relies heavily on renewable electricity and calls for a transition that is: (i) flexible enough to accommodate existing and new electricity end uses and users; (ii) resilient in response to climate change and other threats to electricity infrastructure; (iii) cost-effective in comparison to alternatives; and (iv) just in the face of energy systems that are often the result of—or the cause of—procedural, distributive, and historical injustices. Acknowledging the intertwined roles of technology and policy, this work provides a cross-disciplinary review of how microgrids may contribute to renewable electricity systems that are flexible, resilient, cost-effective, and just (including illustrative examples from Korea, California, New York, the European Union, and elsewhere). Following this review of generalized microgrid characteristics, we more closely examine the role and potential of microgrids in two United States jurisdictions that have adopted 100% renewable electricity standards (Hawai‘i and Puerto Rico), and which are actively developing regulatory regimes putatively designed to enable renewable microgrids. Collectively, this review shows that although microgrids have the potential to support the transition to 100% renewable electricity in a variety of ways, the emerging policy structures require substantial further development to operationalize that potential. We conclude that unresolved fundamental policy tensions arise from justice considerations, such as how to distribute the benefits and burdens of microgrid infrastructure, rather than from technical questions about microgrid topologies and operating characteristics. Nonetheless, technical and quantitative future research will be necessary to assist regulators as they develop microgrid policies. In particular, there is a need to develop socio–techno–economic analyses of cost-effectiveness, which consider a broad range of potential benefits and costs.


Introduction
The climate crisis calls for rapidly transforming the global energy system. As described by the Intergovernmental Panel on Climate Change (IPCC), model pathways for limiting global warming to 1.5 °C indicate that global net anthropogenic carbon emissions will need to decline by about 45% from 2010 levels in the next decade, and reach net zero by around 2050 [1]. Much of these emissions reductions must come from the energy sector, with a particular focus on electricity. The IPCC's successful model pathways are characterized by energy demand reductions, decarbonization of energy systems, and electrification of energy end uses such as transportation and heating. Professor Leah Stokes has

Flexibility and Modularity
The potential flexibility of microgrids is often explained via their potential application to a variety of on-grid or off-grid use cases. These include examples such as electrifying remote or underserved communities, serving critical loads, and developing campus microgrids. However, flexibility characterized solely by end use paints an incomplete picture of microgrid potential.

Modular Grid Planning
Van Nostrand noted that in the grid-connected context, microgrids can offer an alternative to typical electric utility investments and grid planning: Because the optimal size for additions of nuclear, coal, and natural gas-fired generating stations under the traditional utility-scale central generating station model is fairly large, investments by utilities in new generating capacity are said to be 'lumpy,' or available only on a substantial scale. This large scale contrasts sharply with the more steady and smooth growth in demand typically experienced by retail electric utilities. As a result, the resource additions under the traditional utility-scale model often result in a short-term mismatch between loads and resources [8].
This mismatch between resource additions or subtractions, in comparison to growing or shrinking load, can also lead to other imbalances, such as between investments by a regulated utility and charges passed along to consumers under the "used and useful" standard commonly used by regulators to evaluate utility investments [8]. Unless traditional large-scale fossil fuel-fired resources are replaced by decarbonized resources on a uniformly 1:1 basis, the same mismatches can be expected during the forthcoming energy transition.
Moreover, the problem of lumpy investment and deployment is not limited to generation resource additions; transmission capacity is emerging as a key barrier to deploying large-scale renewable energy projects. In the United States, for example, some have identified a need for significant new transmission infrastructure in order to connect coastal population centers with wind resources that are concentrated in faraway Midwest and Plains states [9]. Related proposals, such as a "North American Supergrid," face barriers to deployment, including potential (i) opposition from communities asked to host transmission infrastructure without directly benefitting from the energy it carries, (ii) transjurisdictional wrangling between state governments and the federal government, and (iii) controversial decisions about ownership and eminent domain [10]. These questions about new transmission capacity questions are not unique to the United States. For example, a "Northeast Asian Super Grid" concept would create transmission links between the grids of China, Mongolia, Japan, Korea, and possibly Russia, perhaps in conjunction with largescale renewable energy production in the Gobi Desert [11,12]. A European "megagrid" concept similarly seeks to take advantage of geographic diversity with a European Unionwide interregional sharing of large-scale renewable generation, involving several thousand kilometers of new interregional transmission capacity [13]. Much like in the United States' example, one can envision substantial jurisdictional hurdles for these transmissionheavy concepts, particularly in relation to international borders such as between North Korea and South Korea.
Microgrids and other forms of distributed infrastructure present a flexible alternative, to offset some of the need to expand the grid via new large-scale generation and transmission capacity. Microgrids can deploy distributed generation, load management, and/or ancillary grid services in more modular quantities, nearer to load, and in response to the marginal needs of a grid operator. Moreover, scaling generation to directly match the needs of customers can reduce contingencies and the resulting need for grid-scale reserves [4]. "Reducing demand-both in the long term and by shifting load away from peak-can have the same effects" [4]. Microgrids may enhance this type of demand management, by fostering a network effect and aggregation opportunities among participants.

Flexibility and Decarbonization
With respect to their wide range of potential topologies and end uses, the flexibility of renewable microgrids creates at least one challenge for the task of evaluating their role in a decarbonized electricity system: what is the emissions impact of a microgrid? Certainly, the emissions impact of specific technologies can be evaluated utilizing lifecycle analyses. To the extent that renewable microgrid generation resources displace fossil fuelpowered generation, the resulting lifecycle emissions reductions can be substantial [14][15][16][17]. However, microgrids conforming to the U.S. Department of Energy definition (i.e., those that can island from the grid) are also likely to incorporate energy storage systems. Here, the lifecycle emissions associated with manufacturing, transportation, and installation of storage technologies must be considered alongside operational characteristics. These might include questions about: how the storage is deployed (e.g., is it utilized during day-to-day operation, or solely as a backup resource in islanded mode?); how it is charged (e.g., from renewable generation, or from fossil fuel-fired generation?); whether the use of stored energy creates significant round-trip inefficiencies associated with charging and discharging; and the lifetime of a particular energy storage technology [18].
Perhaps due to the complexity of these questions, microgrids have been the subject of relatively few studies on lifecycle emissions, focused largely on solar-powered remote microgrids [19][20][21][22]. These studies generally conclude that such microgrids "cause significantly less climate change impacts compared to other electricity generation technologies" [19]. Papageorgiou et al. extended this type of lifecycle analysis to a grid-connected battery-backed microgrid in Sweden [19]. Reinforcing the complexity described above, their results indicate that the emissions impact of a microgrid is highly dependent on the marginal source of electricity in the absence of the microgrid. Thus, they conclude that solar microgrids can contribute to decarbonization in areas where the carbon intensity of other generation resources is high, but not in areas with an abundance of low-carbon resources. This observation is highly relevant in the context of considering 100% renewable energy systems, where one may assume that alternative marginal resources will-eventuallyconsist solely of relatively low-carbon renewables.
However, the marginal resources evaluated by Papageorgiou et al. did not include grid-based storage resources. It is safe to assume that many or most 100% renewable electricity systems will incorporate grid-based storage. Raugei et al. quantified the lifecycle emissions associated with adding lithium-ion batteries to a ground-mounted grid-based photovoltaic system under several storage configurations and battery chemistries, and compared the results to fossil-fuel fired generation [23]. They concluded that "broadly speaking, results for all conventional thermal generation invariably indicate over one order of magnitude higher greenhouse gas emissions … generation with respect to all [photovoltaic] systems, regardless of the amount of storage" or the type of battery. Although further research is warranted to evaluate the lifecycle emissions under a wide range of microgrid topologies and operational modes, this result suggests that storage-backed microgrids can indeed play a role in decarbonization.

Resilience
Early microgrids were not developed as a tool for climate mitigation, but rather as a tool for increasing resilience to various vulnerabilities. For example, medical facilities, military facilities, and other energy consumers seeking enhanced resilience have long utilized microgrid architectures to host emergency generating capacity, such as diesel generators [7,24,25]. More modern microgrids can retain this resilience while also deploying decarbonized renewable generation paired with local energy storage. To the extent that climate adaptation involves, in part, increasing the resilience of electricity infrastructure, this dual deployment of climate mitigation capacity and climate adaptation capacity is the recipe invoked by the IPCC and others: "Climate change has started to disrupt electricity generation and, if climate change adaptation options are not considered, it is predicted that these disruptions will be lengthier and more frequent. Adaptation would both secure vulnerable infrastructure and ensure the necessary generation capacity" [8,[26][27][28].
2.2.1. Physical Resilience 2012′s "Superstorm" Sandy illustrates the potential resilience capacity of microgrids [8]. An unprecedented fourteen-foot storm surge swamped grid infrastructure across population centers in the eastern United States, leaving more than eight million utility customers without power-some for weeks. After scrambling to replace thousands of utility poles, thousands of transformers, and hundreds of miles of cable, utilities in the region sought billions of dollars in customer rate increases to apply storm-hardening measures to the grid. In response, a variety of non-governmental organizations and experts advocated for a more deliberate evaluation of the role of microgrids and distributed generation in adding resilience to the region's electricity grid. They noted that many educational institutions, housing communities, hospitals, data centers, and other facilities hosting island-able distributed generation were able to maintain power during the storm's disruptions.
Despite its "Superstorm" tagline, Sandy was not an isolated incident. 2020′s Tropical Storm Isaias led to outages for 900,000 New York utility customers [29]. After an investigation of several utilities' storm response, New York's utility regulator (the New York Public Service Commissions, NYPSC) noted that "more so than any previous time, New Yorkers are depending on essential electric services as a foundation for managing their lives during the ongoing global coronavirus pandemic" [30]. In an order requiring several utilities to show cause why the NYPSC should not seek court-imposed or administrative penalties against the utilities, the commission asserted that "the dramatic and lengthy electric service failure that [NYPSC] Staff observed as a result of [Tropical Storm] Isaias suggested that some electric service providers did not fully appreciate the basic need for safe and reliable electric service." Among other apparent violations, the commission identified failures to contact some "life support" customers (those who require electrically operated machinery to sustain basic life functions), and refer them to emergency responders. "Recognizing prior instances where [the utilities'] storm response had fallen short of legal requirements," the commission noted that if it classified these infractions as repeated violations, the commission would commence a proceeding to revoke or modify the utilities' authorization to operate. This extraordinary potential legal remedy reflects the importance-sometimes life-sustaining importance [31]-of electric grid resilience.
On the other side of the United States, California's grid experienced another form of natural disaster, via wildfires intensified by climate change [32] and, in several notable cases, caused by grid infrastructure [33]. In June 2020, California's Public Utilities Commission adopted short-term actions intended to accelerate the interconnection of microgrids and other "resiliency projects" in advance of the upcoming wildfire season [34]. This order required the state's large investor-owned utilities to "(a) develop and implement standardized, pre-approved system designs for interconnection of resiliency projects that deliver energy services during grid outages; (b) develop and implement methods to increase simplicity and transparency of the processes by which the utilities inspect and approve a project; and (c) prioritize interconnection of resiliency projects for key locations, facilities, and/or customers." The decision also modified net energy metering (NEM) tariffs to allow storage devices to charge from the grid in advance of wildfire threats, and to remove size limits for NEM-paired storage projects.
A microgrid-focused response like California's is supported by technical and theoretical analyses of the microgrids impacts on grid fragility, survivability, and recovery [26]. Liu et al., for example, concluded that "microgrids represent a key component in power grid for improving the grid resilience" (with the caveat that additional work is needed to evaluate the sensitivity of grid resilience on the number of microgrids operating on a system) [35]. Hussain et al. noted other studies evaluating how microgrids can add resilience as a local resource, a community resource, and/or a black-start resource for the wider grid [36]. Syrri et al. provided a framework for assessing the reliability impacts of grid-connected microgrids, concluding that microgrids can offer reliability benefits to microgrid participants and to energy consumers on the wider grid [37,38]. Illustrating the multifaceted potential for microgrid resilience attributes, Strbac et al. described how microgrids might even play a role in increasing the resilience of the European megagrid concept described above [13].

Digital Resilience
Of course, natural disasters are not the electric grid's only vulnerability. Cyber security, for example, is an increasing concern for all infrastructure in the digital age, including microgrids [39,40]. Qi et al. noted that the "threat of cyber-based attacks targeting the … energy sector, and in particular the electric power grid, is growing in number and sophistication" [41]. They (and others) have proposed a multilayered framework for modeling, preventing, detecting, and responding to cyber threats, in response to the concern that increasing penetration of distributed energy resources will result in a proliferation of devices and access points outside a utility's direct administration and thus expand the vulnerable "attack surface." It is not clear, however, that distributed resources nested within networked microgrids are inherently more vulnerable to cyber threats than existing centralized grid architectures. This is especially apparent when one considers the traditional role of microgrids that were developed especially to serve critical loads in the face of these types of vulnerabilities. Veitch et al., for example, described a microgrid security architecture for military microgrids in the United States that segments control systems based on functional necessities, physical locations, and/or security concerns [42]. They asserted that this type of isolation can minimize malicious opportunities, provide good locations for intrusion detection, and improve network performance. This approach appears to operationalize the type of potential microgrid security benefits described by Qazi and Young, who postulated that solar-based microgrids are less vulnerable to cyberattacks because of their ability to island from the main grid [43].

Levelized Cost of Energy
The narrative of inherently "expensive renewables" is eroding. For example, asset management and financial advisory firm Lazard's most recent study of the levelized cost of energy (LCOE) reported that the unsubsidized cost of new utility-scale wind and solar generation is generally lower than the cost of new fossil fuel-fired generation-particularly when accounting for the sensitivity of fossil fuel-fired generation to fuel prices [44]. Lazard noted, of course, that LCOE does not allow for a fully direct comparison between all types of generation, as it does not account for potential social and environmental costs, geographic distribution, dispatch characteristics, or reliability-related considerations. Nonetheless, the LCOE analysis does illustrate the trend that decreasing capital costs, improving technologies, and increased competition, among other factors, are driving down the price of renewables. Moreover, Lazard reported that combining distributed solar generation with energy storage at the microgrid scale (i.e., residential, or commercial and industrial applications) results in a lower levelized cost of storage compared to standalone storage at the same scale.
Microgrids and microgrid-scale storage can also contribute other energy services and functions, such as demand response capacity, frequency regulation, resource adequacy, spinning reserve, and backup power during grid outages [45,46]. Value streams like this, previously largely served by fossil fuel-fired generation, are key components of 100% renewable electricity systems. Thus, while Lazard and others report that levelized costs of renewables at the microgrid scale are presently higher than costs at larger scales, it appears possible that microgrid-scale renewables may be able to capture a variety of these other potential value streams, while continuing to benefit from the same factors driving down costs at other scales (e.g., improving technologies and increased competition).

Considerations beyond Levelized Cost of Energy
Various authors have identified the types of costs and benefits that can help to define this broader view of microgrid cost-effectiveness. Brown asserted that with an appropriate methodology to value grid-edge resources, "microgrids employing multiple energy management technologies can simultaneously provide multiple dynamic objective functions" [4]. These include the ability to adjust generation and load, to shape an aggregate profile, to shift load, and to locate generation closer to load-strategies that can help moderate power prices and manage grid congestion [4]. Resilience is another relevant value stream. Anderson et al. concluded that monetizing the resilience value of renewable energy systems at the building and campus scale (i.e., relevant to microgrids) is a multibillion-dollar opportunity in comparison to existing energy backup systems [47]. These concepts reinforce the idea that traditional notions of cost-effectiveness (e.g., LCOE) do not capture all relevant costs and benefits relevant to the current electricity transformation. Table 1 identifies some of the cost-effectiveness considerations that may be particularly relevant to microgrids. However, utility business and regulatory models have not yet resolved how to monetize all of these costs and benefits. For example, it remains challenging to value resilience in the context of evaluating the prudence of proposed gridhardening investments [48][49][50]. Cost-effectiveness also invokes comparison to alternatives. In the current energy transition, those comparisons must go beyond the fossil fuel generating resources identified in Lazard's LCOE analysis. Larsen et al. noted that storm hardening of existing grid infrastructure would cost trillions of dollars [51]. This is relevant to assessing the costeffectiveness of microgrids if they contribute to physical resilience. A robust view of energy costs in the context of the climate crisis should also explicitly consider the cost of carbon sequestration or other mitigation strategies for GHG-emitting alternatives. Related to the potentially troublesome task of combining large-scale renewable generation with large-scale transmission infrastructure, Bronin argued that the resulting "energy sprawl costs space, money, and energy itself" [52]. These costs are related to the balance of costs and benefits in existing land uses and land-use policies, the expensive need to coordinate across jurisdictional and ownership boundaries, and the energy losses associated with transmission.
Polly et al. and Saleski et al. evaluated the potential cost-effectiveness of zero-energy districts [53,54]. These are conceptualized as parts of cities designed to balance energy consumption and generation, via district energy systems that coordinate energy generation and demand, energy storage, and waste heat, etc., across multiple buildings. Compared to standalone net-zero energy buildings, these urban districts may be able to improve energy cost-effectiveness by leveraging characteristics such as load diversity, scale, and coordinated participation in markets for ancillary grid services [53,54]. This is particularly relevant, because (i) governments are implementing net-zero energy building standards for new or rebuilt buildings, (ii) population is shifting to urban areas, (iii) buildings account for nearly 40% of energy-related GHG emissions, (iv) the growth in building floor area is outpacing population growth [55][56][57]. Microgrids-like zero-energy districts-can also coordinate generation and load, and may also enjoy other similar characteristics such as load diversity and coordinated participation in markets for ancillary services. Thus, it seems worthy for future research to explore the role of microgrids in urban zero-energy districts, and to further quantify the cost-effectiveness benefits of these coordinated systems.
The above examples illustrate the tangled role of technology and policy in evaluating the cost-effectiveness of various components in the energy transformation. Microgrids are likely to play a substantial role in a decarbonized energy system only if mechanisms such as electricity tariffs (applicable to the provision and consumption of energy and energy services by microgrids), prudence reviews of utility investments, building standards, and land-use policies are able to assess or define the value of microgrid characteristics. Hatziargyriou et al., quantifying the economic, environmental, and operational benefits of microgrid penetration in Greek networks, similarly argued that "microgrids will only mature as a viable market alternative for consumers and utilities when all the benefits provided by a particular microgrid are accounted for and credited to its owners" [58]. Yet the characteristics (and resulting benefits and costs) of a microgrid are, at least in part, dependent on the technology mix incorporated into the microgrid. In turn, communities and broader energy systems are only likely to invest in a technology if it appears that the value outweighs the costs.
While prior works acknowledge this interdependency between policy and technology, the transition to 100% renewable energy prompts re-evaluation. For example, a thorough review of microgrid economics by Milis et al. concluded that "where the most often viable reported system configuration is concerned, the current scientific consensus is that [combined heat and power]-powered microgrids are the most economically viable type of microgrid under a wide range of policy interventions, with renewable powered microgrids only viable in fringe cases, making the often-heard claim that microgrids will allow for greater penetration of renewable sources highly suspect at best" [59]. In the context of the growing number of jurisdictions planning for 100% renewable electricity systems, this "fringe case" is the norm, yet that scenario is the subject of only one study reviewed by Milis et al. Accordingly, the review acknowledged that "a lot of research attention has been devoted to the topics of carbon taxation and TOU-tariffs," with "other subjects such as tax incentives or command and control policies only receiving moderate research attention." Furthermore, "concerning the objective of outlining any research gaps, [the review showed] that the exploration of the impact of tariff systems other than TOUpricing on the optimal configuration of a microgrid hasn't received research attention." In addition, the economic impact of other important policy-influenced microgrid characteristics, such as infrastructure scale, location, and effects on land use, should also be accounted for. An even broader socio-techno-economic analysis, including these elements and more (e.g., the impact on jobs), will also be required to account for how policy and technology decisions are succeeding or failing to promote energy justice (discussed below).
Gaps in the evaluation of microgrid cost-effectiveness appear to be long-lived. Discussing policy-making for microgrids in 2008, Marnay et al. noted that "microgrids bring to the fore the issues of waste-heat-driven cooling, on-site energy storage, and heterogeneous [power quality and reliability], which [were] all relatively uncharted areas of engineering-economic analysis" [60]. It appears that, perhaps as much as any other potential component of 100% renewable energy systems, microgrids suffer from a "chicken or egg" problem in the realm of assessing cost-effectiveness in the midst of co-evolving policies and technologies. This suggests that an analysis of the economic impacts of microgrids in 100% renewable energy systems, considering a wide range of potential characteristics, costs, and benefits, is a useful area for further inquiry.
Remote and island microgrids may provide fertile ground for this line of inquiry. For example, Gasa Island and Gapa Island in the Republic of Korea are each the site of microgrids that incorporate renewable generating capacity. Gapa's microgrid was developed with two wind turbines (250 kW) and almost fifty solar panels (174 kW total), supplemented by a 3.85 MWh battery system and diesel generators [61][62][63][64]. The subsequent renewable component of the islands electricity has reportedly ranged from approximately 40% to 80%, and as of 2018, the island could self-sustain for seven days. With the battery capacity doubled, this mode of operation is expected to last up to 25 days. In the meanwhile, the microgrid project has been associated with a substantial decrease in local energy costs. Before the project, the average monthly electricity bill of each household on Gasa was around 120,000 to 130,000 won; this reportedly dropped to approximately 20,000-25,000 won with the microgrid project [63].
Similar results have been reported for a microgrid project on Gasa Island, serving approximately 165 households, lighthouses, waterworks, and military radar facilities. This microgrid utilizes four 100 kW wind turbines, four photovoltaic installations totaling approximately 320 kW, and three diesel power units totaling 450 kW [62,65]. Excess energy is stored in a 3 MWh battery system, which enables full-day electricity consumption throughout the whole island if fully charged [66]. After installation, the consumption of diesel dropped substantially [67], and Kim and Mathews reported a 200 won/kWh savings compared to diesel generation [65]. It would be interesting for future research to explore how the reported savings on Gasa and Gapa relate to the levelized cost of energy, and how broader treatment of costs and benefits might be compared to the scale of the reported savings.
Korean islands have long been identified as prime candidates for microgrids [68], and may be particularly useful sites to study these cost-effectiveness issues in the context of 100% renewable energy standards. The Jeju province, home to Korea's largest and most populated island (and also home to the much-smaller island of Gapa), adopted a 100% renewable energy target (2030) in 2012 [69], making it one of the earliest jurisdictions to adopt such a policy. This initiative was expected to act as an important reference source for South Korea's broader energy transition, creating a test bed for modern grid technology [65]. With the 2020 adoption of a South Korean Green New Deal policy targeting carbon neutrality for the entire country by 2050, this test bed will naturally inform Korean decarbonization efforts beyond remote island settings [70]. Indeed, Korea has already developed microgrids in a number of urban settings [61]. Similarly, Hawai'i and Puerto Rico (discussed below) have each adopted 100% renewable electricity standards, and both jurisdictions include a mix of remote island settings and urban areas. This makes them suitable locales to investigate the role of renewable microgrids.

Examples of Energy Justice Principles Embedded in Policy
International, national, and subnational policies have called for the renewable energy transition to incorporate justice [71]. The Paris Agreement, for example, incorporates justice principles that are applicable at the international and national scale: "[The Parties will implement the Agreement] to reflect equity and the principle of common but differentiated responsibilities and respective capabilities, in the light of different national circumstances" [72]. The Agreement also references justice principles more applicable at community scales, claiming to account for "the imperative of a just transition of the workforce and the creation of decent work and quality jobs in accordance with nationally defined development priorities." It also acknowledges that climate action should "respect, promote and consider their respective obligations on human rights, the right to health, the rights of indigenous peoples, local communities, migrants, children, persons with disabilities and people in vulnerable situations and the right to development, as well as gender equality, empowerment of women and intergenerational equity." With these words, the Agreement "became the first multilateral environmental agreement to contain an explicit reference to human rights, albeit in its preamble" [73].
Recent European Union (EU) policy-making also provides an example of energy justice-relevant policy. The 2019 EU electricity market directive includes a variety of provisions related to small consumers in liberalized European energy markets, with a general approach self-described as "competitive, consumer-centred, flexible and non-discriminatory" [74]. (Here, the phrase "non-discriminatory" appears to be largely focused on avoiding barriers to market entry, and avoiding cross-participant subsidization, rather than non-discriminatory in other senses of the phrase.) Among other provisions, the directive provides for "citizen energy communities" to participate in energy markets. These are legal entities: (a) that "are based on voluntary and open participation and is effectively controlled by members or shareholders that are natural persons, local authorities, including municipalities, or small enterprises;" (b) whose "primary purpose [is] to provide environmental, economic or social community benefits to its members or shareholders or to the local areas where it operates rather than to generate financial profits;" and (c) who "may engage in generation, including from renewable sources, distribution, supply, consumption, aggregation, energy storage, energy efficiency services or charging services for electric vehicles or provide other energy services to its members or shareholders." Although it remains unclear exactly how these communities will participate in various energy markets, Mostert and Naude asserted that this concept has a bearing on energy selfsufficiency: "The implication of this provision may be that more traditional energy supply companies have to lower their prices if they wish to discourage such local communities from generating their own electricity" [75]. Roggenkamp and Diestelmeier also discussed the extent to which the EU directive may influence energy justice within member states, with a particular focus on how provisions related to energy poverty and the protection of small electricity customers are translated into the energy policy of member states [76]. They described that, although the directive recognized a role for enhanced participation by consumers (i.e., prosumers), uncertainty and disparity remain across member states in terms of defining vulnerable classes of customers and consumer protections could. They cautioned that this could result in a two classes of prosumers: "one which is able to 'take ownership of and benefit from new technologies' and another category which is not able to do so and remains vulnerable to or even suffers from energy poverty." An example of local energy policy that references justice-relevant components can be found in the Seoul Metropolitan Government's 2012 "One Less Nuclear Power Plant" (OLNPP) initiative [77]. Ahn described energy justice as a "pillar" of the initiative, insofar as it was framed around the need for Seoul to become more energy self-sufficient, rather than sourcing power from outside the city via generation and transmission infrastructure that entails "excruciating social conflicts" including "conflict over the construction of high-voltage transmission cable towers in Miryang in southern Korea" [78]. Byrne and Yun described that the policy also sought to "remove unequal burdens among members of a society to enjoy needed energy services," including via the utilization of energy cooperatives and micro-scale solar generation that could be utilized on apartment verandas in Seoul's dense urban setting [79]. They also described that the initiative characterized this commitment using terminology "with the approximate meaning of 'energy fairness,' 'energy equity' and 'energy justice' employed by researchers and some countries in characterizing a social condition or metric for unaffordable energy services for sizable segments of a society." Lastly, the initiative considered citizen participation. In some ways, this appears focused on individual participation in energy efficiency measures. However, the initiative also touted citizen input into policy-making, including the establishment of a Citizens' Commission comprised of "19 reputable figures from civic groups, the business & media arena as well as religious, educational and cultural sectors" [77]. While it is not clear whether this commission advances the voice of marginalized portions of the citizenry, Ahn described this commission as responsible for "leading the policy paradigm shift from an energy saving city to an energy production city, "determining policy directions for the OLNPP initiative, and "reviewing the OLNPP initiative action plans and their revisions and making overall adjustments" [78].

The Potential Role of Microgrids in Operationalizing Energy Justice Principles
These examples illustrate that energy justice can take on a wide variety of contextualized meanings in policy at various scales. None, however, demonstrate that energy policies have yet succeeded in operationalizing the three core energy justice principles identified in Figure 1: procedural justice, distributive justice, and restorative justice [5]. In this framework, energy justice requires that decision-making processes must: (i) fairly and competently incorporate marginalized perspectives and communities (procedural justice); (ii) equitably distribute the benefits and burdens of generation, transmission, distribution, consumption, and other elements of energy systems (distributive justice); and (iii) repair past and ongoing harms caused by energy systems (restorative justice) [5,[80][81][82].
Recent literature evidences a proliferation of academic work on energy justice concepts, yet reinforces the conclusion that much work remains in operationalizing these principles. Welton and Eisen, for example, explored various distributive and procedural justice opportunities and challenges in the energy transition, and identified a "paucity" of data on clean energy's justice implications [83]. They highlighted data gaps on issues such as the definition and distribution of "clean energy jobs," and they identified a need to pair data on energy poverty and its correlation (or not) with policies related to distributedscale solar, grid modernization, community solar, and other topics. With respect to the distributive justice implications for siting energy infrastructure, they sensibly noted that "much of the community impact of wind and solar energy turns upon scale-that is, the size of a proposed installation," and they noted the potential for large-scale renewable energy infrastructure, and its necessary transmission infrastructure, to create a rural/urban energy justice divide. Others, such as Finley-Brook and Holloman, and Zhou and Noonan, have similarly recently identified empirical gaps in understanding the justice implications of energy policy [84,85]. Despite efforts like these to systematically identify energy justice questions, challenges, and opportunities, the potential role of microgrids in operationalizing energy justice principles has received relatively less attention. Wolsink argued that distributed generation microgrids can indeed play such a role, with respect to each of the three energy justice tenets [86]. Identifying microgrids as sociotechnical systems, Wolsink explained that microgrids offer an opportunity to reorganize traditional roles in electricity production and consumption, particularly via the new social relationships within the boundaries of a microgrid (e.g., participation, ownership, management, etc.). Without specific reference to microgrids, Banerjee et al. similarly reviewed ways in which community-scale renewable models can promote justice "by virtue of inclusive participation, collective ownership, and community empowerment" [87]. Welton observed that "the history of electrification counsels that our most successful grid experiments in terms of equity and empowerment may come from focusing on more collective forms of grid participation. Thus, regulators might pay particular attention to programs like community solar and microgrid formation for the community-scale participation that they embody" [88].
Others have evaluated a substantial potential role for microgrids in promoting energy access via rural electrification [89][90][91]. Venkataramanan and Marnay pointed to micro-utility models pioneered in Bangladesh as "evidence of the relative ease with which grass-roots solar electrification projects can be carried out without heavily subsidized large capital development assistance, when appropriately integrated with community economic development. Similar applications of renewable energy technologies such as micro-hydro, photovoltaic cells, and small-scale wind turbines are well developed and have been deployed widely in the developing world" [92]. This suggests that the microgrid characteristic of flexibility and modularity, discussed above in the context of grid planning and infrastructure, may also be a beneficial characteristic with respect to energy justice, in the form of universal energy access.
The ability of microgrids to account for hyper-localized energy and development needs may render them particularly relevant in the context of island electrification. Veilleux et al., for example, analyzed and cited the potential for microgrids to electrify island communities in places such as Thailand, the Pacific, and Indonesia (where "more than 50% of the unelectrified are believed to live on islands") [93]. Bertheau, considering the electrification of island communities in the Philippines, concluded that "100% [renewable energy] systems are a suitable option for electrification and could allow a high energy autonomy and little operational costs" [94]. Moreover, this analysis noted that 100% renewable energy designs often utilize excess generation capacity, creating an opportunity for economic development around non-critical loads, to utilize otherwise curtailed energy that is available at essentially zero marginal cost.
However, even if one assumes or concludes that microgrids foster collective participation, universal access, and a high-penetration of renewables, this does not inherently answer the challenge of energy justice. Questions remain about who can participate in, own, or manage a microgrid; on what terms; and for whose benefit. Here, the flexibility of microgrid architectures presents a challenge as much as it presents an opportunity. Thus, Wolsink and others also identified the need to update other institutional frameworks necessary to organize multiple microgrids-or other renewable energy infrastructure-within a larger energy system (e.g., tariffs, access to capital, etc.) [86,95]. Schnitzer et al., discussing best practices for microgrid development based on seven case studies, identified the following additional "critical factors" for successful microgrid development: "tariff design, tariff collection mechanisms, maintenance and contractor performance, theft management, demand growth, load limits, and local training and institutionalization" [96].

Hawai'i
In 2015, Hawai'i adopted a 100% renewable energy standard that will require regulated electric utilities to achieve a 100% renewable portfolio standard (RPS) no later than 2045 [97]. Utilities in the state have exceeded the law's most recent interim RPS target (30% by 2020). The state's largest utility, the Hawaiian Electric Companies, reported a consolidated 2020 RPS of 34.5% across its entire service territory [98]. Individual island systems have achieved higher RPS metrics, with the island of Hawai'i reporting 43.4% and the county of Maui (comprising three non-interconnected islands) reporting 50.8% for 2020. Because the law calculates the RPS as renewable energy generation in proportion to electricity sales, these RPS values are approximately 4% to 11% higher than "total" renewable energy, which more simply accounts for renewable generation in proportion to total generation [99]. Distributed generation (predominantly rooftop solar) accounted for nearly half of Hawaiian Electric's 2020 renewable generation, while utility-scale wind and solar power accounted for approximately one-fifth and one-sixth, respectively. Biomass (including municipal solid waste combustion), biofuels, hydropower, and geothermal power accounted for the remainder of Hawaiian Electric's renewable generation in 2020.
The island of Kaua'i's electric utility (Kaua'i Island Utility Cooperative, KIUC) has not, as of this writing, reported its 2020 RPS. It reported 56.5% for 2019 [100]. KIUC also reported that it "is now routinely running on 100 percent renewables for five hours or more on sunny days" [101]. In addition to distributed solar, hydropower, and biomass, the KIUC system also utilizes utility-scale solar, including battery-backed utility-scale solar.
While each utility in Hawai'i is actively implementing resource plans to achieve 100% renewables, ongoing iterations of those plans continue to evaluate options for specific resource additions and retirements. Similarly, the academic literature continues to model and evaluate technical pathways to 100% renewable systems [102][103][104][105][106][107][108][109][110][111]-including an exchange between Brown et al. and Heard et al. debating whether the feasibility and viability of 100% renewable systems had yet been demonstrated as of 2017 and 2018 [112,113]. Modeling by Imelda et al. concluded that a 100% renewable system satisfying Hawai'i's 100% RPS law is "surprisingly affordable" and improves welfare in comparison to a fossil fuel-based system even without considering the cost of pollution [104]. Microgrids do not appear to play a substantial role in these works, although related concepts have been reviewed in the context of 100% renewables, such as Eras-Almeida and Egido-Aguilera's review of renewable island mini-grids and Weinand et al.'s review of decentralized autonomous energy systems [114,115].
Even before Hawai'i adopted its 100% RPS, regulators in the state envisioned a role for microgrids in supporting the transition to renewable energy. In 2014, the Hawai'i Public Utilities Commission (HPUC) issued a landmark set of inclinations on the future of the state's energy system: Technological innovation is supporting the development of integrated energy districts that aggregate pockets of load and generation resources, which can disconnect and reconnect to the main grid in times of emergency. A subset of this aggregation concept is sometimes described as a microgrid. Several microgrid demonstration projects are underway in Hawai'i and large energy customers are investigating the development of these systems to meet their energy needs. As the island electric systems evolve, the utilities' transmission system planning needs to address the potential development of integrated energy districts and, as the technology matures, these systems will need to be evaluated as potential non-transmission alternatives to expansion of the transmission system [116].
Today, Hawai'i is home to at least one operational microgrid, based around a 50 megawatt thermal generating station at the Schofield Barracks U.S. Army installation [117]. This generating station is a joint project between the U.S. Army and the investor-owned Hawaiian Electric Companies. It is designed to be fuel-flexible; pursuant to the Army's project requirements, it must utilize fifty-percent biofuel, or three million gallons of biofuel per year, whichever is less [117]. During emergencies, the facility can provide islanded power to military facilities. It is also capable of providing grid services such as: black start capability during outage recovery; fast ramping for frequency regulation; voltage regulation; and inertial response.
Another microgrid is under development at the Natural Energy Laboratory of Hawai'i ("NELHA"), where research on ocean thermal energy conversion and other uses of cold seawater rely on a seawater pumping system, alongside various energy and other research activities. This microgrid will be designed to increase the facility's energy resilience by serving the pumps' critical load via 600 kW of solar generating capacity, 585 kWh of battery storage capacity, existing diesel generating capacity, and automated microgrid management software [118]. The effort is a collaboration that includes LG Electronics, Encored Inc. (a U.S.-based company with Korean ties), Seoul National University, Gwangju Institute of Science and Technology, and the Hawai'i Natural Energy Institute at the University of Hawai'i at Mānoa.
To promote further development of microgrids, the Hawai'i legislature passed Act 200 in 2018 [119]. The legislature asserted that "microgrids can facilitate the achievement of [Hawai'i's] clean energy policies by enabling the integration of higher levels of renewable energy and advanced distributed energy resources;" and that "microgrids can also provide valuable services to the public utility electricity grid, including energy storage and demand response, to support load shifting, frequency response, and voltage control, among other ancillary services." The legislature also noted that microgrid policy had not kept pace with technological capacity: microgrid "development has been inhibited by a number of factors, including interconnection barriers and a lack of standard terms regarding the value of services exchanged between the microgrid operator and the utility." On this basis, Act 200 directed the HPUC to create a microgrid services tariff, designed to standardize microgrid interconnection and to assess the value of microgrid services.
To design the tariff, the HPUC's still-ongoing regulatory docket involves the Hawaiian Electric Companies, the state's Office of Consumer Advocacy, and six entities, including a microgrid industry trade group, an energy consulting business, renewable energy industry trade groups and advocates, an impact investment firm, and two community groups (both groups later withdrew from the docket).
This process has amply illustrated policy gaps that must be resolved in determining how microgrids will support a 100% renewable electrical grid in the state. As an initial substantive step, the HPUC identified a set of seven "preliminary"-yet fundamentalquestions to define the scope of the microgrid services tariff [120]. Table 2 outlines these questions. Although the questions were labeled as "preliminary," it appears that only one was readily resolved; the HPUC adopted the oft-cited U.S. Department of Energy's microgrid definition for use in the context of a microgrid services tariff [6,121]. With respect to microgrid characteristics, the commission's preliminary determination also allowed for microgrids to "have a mixed resource profile," i.e., microgrids are not initially required to operate using 100% renewable generation. In particular, a draft tariff specifies that microgrids operating in islanded mode shall not be included in calculating the utilities' renewable portfolio standard [122]. However, the commission noted that it "expects the inclusion of more renewable resources in the energy portfolio of future microgrids that support the achievement of Hawai'i's [renewable portfolio standard] goals. Those who participate in the microgrid services tariff should be cognizant of the State's broader energy policy goals, and aim to develop microgrids that primarily consist of renewable energy resources" [121]. The details of this future requirement, and the remaining "preliminary" questions, were left for further discussion. To this end, the commission formed two working groups to investigate the issues, with the goal of providing a microgrid services tariff to the HPUC for review and potential approval.
The working group process again illustrates the unresolved state of fundamental policy questions. For example, the groups recommended categorically excluding three types of microgrids from the tariff: (i) utility microgrids (interconnected and compensated via existing regulatory processes); (ii) remote microgrids (not interconnected to the broader grid); and (iii) virtual microgrids (not compatible with the applicable definition of a microgrid) [123]. The exclusion of remote microgrids does not indicate that such grids cannot be developed in Hawai'i, but rather that they will not provide grid services in the manner envisioned by a microgrid services tariff. Remote microgrids will require answers to different policy questions. For example, should remote microgrid operators be regulated as public utilities?

Flexibility and Modularity in Hawai'i's Microgrid Services Tariff
Elements of Act 200 expressly called for a flexible microgrid services tariff. The tariff must accommodate any microgrid ownership and operation model ("Any person or entity may own or operate an eligible microgrid project or projects"). In other ways, it appears that eligible microgrid models will be constrained. The working groups recommended that microgrids with generation capacity above three megawatts should not be eligible for the tariff, if the microgrid utilizes both utility and non-utility infrastructure beyond the point of common coupling (e.g., utility distribution infrastructure). These are designated as "hybrid microgrids" in the draft tariff. The groups reasoned that the potential complexity of such arrangements calls for individually negotiated power purchase agreements, rather than standardized interconnection and compensation terms. This appears to be partially consistent with Brown's view that the "diversity of [microgrid] capabilities cannot be integrated into the grid through a one-size-fits-all, grid-edge resource tariff, but only through valuation of the particular services provided by a particular gridedge resource" [4]. In contrast to this approach, the groups recommended that microgrids utilizing only non-utility infrastructure beyond the point of common coupling should still be eligible for the tariff, even if the microgrid generating capacity is greater than three megawatts. These are designated as "customer microgrids." Figure 2 depicts simple versions of these two categories of microgrids. This distinction between microgrids that use utility distribution infrastructure, versus those that do not, is related to questions about how utilities should be compensated for the use of utility infrastructure in delivering grid services. For example, the value of a microgrid's grid services might be partially offset by the value of utility infrastructure used to deliver those services. One view might suggest that compensating a microgrid for grid services delivered via utility infrastructure will result in ratepayers paying twice for the same infrastructure. Another view might suggest that the purpose of modern utility infrastructure is to enable the delivery of grid services, whether delivered by the utility, or delivered by microgrids or other forms of distributed energy infrastructure.
With respect to a capacity limitation, it is not clear that three megawatts reflects an optimized and universally applicable cap, above which standardized interconnection and compensation terms become impracticable or inefficient. Rather, the Hawaiian Electric Companies indicated that this cutoff was selected based on approximate feeder capacity. The HPUC recently requested that the parties reconsider altering, or eliminating, the cap on project sizes [124]. It is not clear whether many realizable use cases for larger microgrids are likely to be prevalent in Hawai'i (although the 50 megawatt Schofield microgrid suggests that it is possible). The general characteristic of flexibility calls for a tariff that would make it possible, even if unlikely, to develop a larger microgrid using the tariff's standardized terms.
The HPUC and working group also considered whether additional regulatory flexibility was warranted to support demonstration projects. However, it appears that the NELHA demonstration project will be interconnected via a standard interconnection agreement [123], and this concept of regulatory flexibility for demonstration projects has not been otherwise prioritized by the HPUC and working groups.

Resilience in Hawai'i's Microgrid Services Tariff
Act 200 invoked a specific legislative finding that "Hawaii's residents and businesses are vulnerable to disruptions in the islands' energy systems caused by extreme weather events or other disasters," and thus "the use of microgrids would build energy resiliency into our communities, thereby increasing public safety and security." Despite this rationale, Act 200 did not mandate that a microgrid services tariff must monetize or otherwise compensate for the resilience value of microgrids. Rather, the tariff is more broadly required "to provide fair compensation for electricity, electric grid services, and other benefits provided to, or by, the electric utility, the person or entity operating the microgrid, and other ratepayers." This language was modeled on earlier legislation launching Hawai'i's community solar program, and thus does not appear tailored to microgrids [125].
The working groups initially considered the potential scope of these compensated "other benefits," such as resilience. However, such consideration is complicated by a distinction between added resilience for the participants within a microgrid, compared to resilience for the grid at-large. The HPUC ordered that the working groups should prepare draft tariff language that, "as an initial step of development supports resilience of energy services during emergency events and grid outages" [121]. This was later clarified to mean that the priority is to "enable microgrids that can disconnect from the grid to operate in island mode during emergency events or grid outages" [126]. Under this prioritization, resilience is a benefit acquired primarily by the microgrid participants, and thus it is not viewed as necessary for the tariff to compensate for broader grid services. Whether a utility tariff should be used to compensate the potential societal value of resilience within a microgrid remains an open question [123]. For example, would, or should, such compensation promote the development of microgrids that host facilities that can be used as emergency shelters? The working groups have suggested that this issue might be addressed in a future resiliency tariff, reserved for "showing of broad-based benefits for nonparticipants" [126].
The microgrid services tariff must nonetheless address complexities related to a microgrid's islanding function. For example, a working group lead asserted that if a microgrid has a unilateral right to island for non-emergency reasons, this "could be viewed in its logical extension as enabling grid defection, with the microgrid only using the utility system as back-up." A trade group, the Microgrid Resources Coalition, offered the competing opinion that "if a microgrid has no contractual service obligation to the grid, it can island if it wishes to," reasoning that "a sophisticated microgrid should be able to exit or enter parallel operation [with the grid] at neutral load, so as not to cause or exacerbate any problem on the grid." Provisions related to this "anti-islanding" concept are still under revision in the draft tariff. Thus, while this issue is not yet resolved from a regulatory standpoint, it does signal the eventual technical expectations for microgrids utilizing the tariff.
The draft tariff also contemplates monitoring and reporting, which are important elements of resilience for the grid at-large. Hybrid microgrids (i.e., those utilizing both utility and non-utility distribution infrastructure) will be required to provide a secure means of communication between the utilities' SCADA [supervisory control and data acquisition] system and the microgrid controller, and must include a range of variables, such as voltage and power flow at the point of common coupling, reserve capacity, the status of controllable distribution assets inside the microgrid, remaining load-serving duration, and other data points [122].

Cost-Effectiveness in Hawai'i's Microgrid Services Tariff
The draft tariff is viewed as a "portal tariff" that guides microgrid applicants to existing tariffs designed to compensate for electricity or services provided by, or to, the grid at-large. Thus, the HPUC has indicated that it is not necessary for a microgrid services tariff to establish new modes of compensation related to microgrids [127]. Instead, the commission referred this issue to an ongoing docket exploring how distributed energy resources ("DERs") should be compensated (Docket No. 2019-0323), and to potential future discussion on the role of retail wheeling for energy and services. Thus, it does not appear that the adoption of this microgrid services tariff will have a substantial standalone effect on the cost-effectiveness of grid-based electricity or services in Hawai'i. The effect of the microgrid services tariff, if any, will be focused on the initial step of clarifying interconnection requirements. Any impact on cost-effectiveness in Hawai'i will be determined, in the immediate future, primarily by the costs or savings achieved by a microgrid's direct participants. If clarifying interconnection requirements promotes new microgrid development, and if those microgrids participate in various other tariffs (e.g., DERs, demand response aggregation, etc.), there may be a secondary effect on cost-effectiveness via those other tariffs. The Microgrid Resources Coalition (represented by C. Baird Brown, cited earlier in this review) critiqued this approach. They asserted that the draft tariff failed to streamline microgrid interconnection, and made only a "nod" to microgrid compensation by making them eligible for the same programs as other resources-without properly considering how to "compensate microgrids for services they can uniquely provide" [128].

Energy Justice in Hawai'i's Microgrid Services Tariff
It does not appear that the microgrid services tariff will include a focus on energy justice. Act 200 and the resulting draft tariff allow for ownership by "any person or entity." However, there is nothing in the draft tariff that would favor or promote microgrids that otherwise promote equity in a distributive, procedural, or restorative sense, particularly in the absence of a means to compensate microgrids for specialized value streams. Fundamental barriers such as access to capital will remain a hurdle for underserved communities.
Nonetheless, it is possible that some communities may find that the tariff enables interconnection for microgrids that serve specialized needs. For example, the Hawaiian Homestead community of Kailapa on the island of Hawai'i has created a Community Resilience Plan [129]. As its first priority, the community has identified the need for secure access to fresh water. This need is sharpened by rising costs (residential water costs are projected to increase by 400 percent over ten years), the insecurity of water acquired from private entities (potential loss of access with two years' notice), and limited access (water is available for residential purposes only, curtailing agricultural opportunities). Potential solutions include a system of water storage and transmission that is owned and operated by community residents. The Resilience Plan also identifies the need for "renewable energy projects to create self-sufficiency and economic opportunities for [community] members." Combining the potential need for energy capacity to serve a new water system, along with this push for additional energy self-sufficiency, the community has apparently considered energy solutions such as pumped storage hydropower [130]. Although the community is grid-connected, it is relatively remote. It is quite easy to envision that the planned approach to resilience could utilize a community microgrid. It is possible, although far from certain, that the microgrid services tariff could enable or promote interconnection of such a microgrid.

Puerto Rico
Hawai'i's Act 200 mandated that the HPUC must look to other jurisdictions for guidance on promoting microgrids-with specific reference to Puerto Rico's efforts to address energy resilience. Puerto Rico's Regulation 9028 on Microgrid Development was promulgated in 2018 by the Puerto Rico Energy Commission (now the Puerto Rico Energy Bureau, "PREB") as part of the Government of Puerto Rico's strategy for rebuilding and strengthening the electric power system in the aftermath of 2017′s Hurricanes Irma and Maria [131]. These storms were associated with excess mortality totaling in the thousands [132] and an electricity outage that continued to affect half the population after seven weeks [133]. This outage stretched to six months for some communities [134], and the utility did not report reconnecting the last of its 1.5 million customers until eleven months after the storm [135].
As of September 2019, hundreds of renewable microgrids had reportedly been deployed in Puerto Rico to increase energy resilience for schools and other critical loads [136]. These microgrids are set to be a component of the transition to 100% renewable electricity; in 2019, Puerto Rico adopted the Puerto Rico Energy Public Policy Act, setting a 2050 target date for its 100% renewable portfolio standard [137]. However, even before that adoption, Regulation 9028′s detailed framework for microgrid regulation referenced the decarbonization value of microgrids, in addition to their value in avoiding loss of power at critical facilities, and other benefits.

Flexibility and Modularity in Puerto Rico's Microgrid Services Tariff
Regulation 9028 allows for a range of microgrid models, described in three categories: personal microgrids (owned by no more than two persons and producing energy primarily for consumption by its owner(s)); cooperative microgrids (jointly owned by cooperative members through formal or informal organization, so long as no single member owns more than a 35% interest, with a primary purpose of supplying energy services or grid services to its cooperative members); and third-party microgrids (a catchall for anything other than a personal or cooperative microgrid, owned and operated by any person for the primary purpose of engaging in the sale of energy services and grid services to any customer).
The regulation contemplates that personal and cooperative microgrids may sell excess energy or other grid services to third parties besides the owner(s) of the microgrid and Puerto Rico Electric Power Authority (PREPA)-the government-owned and -operated monopoly utility. Unlike Hawai'i's proposed three-megawatt cap for some microgrids, Regulation 9028 appears to allow microgrids of any size, with the caveat that operators of third-party microgrids greater than one megawatt will be considered regulated Electric Service Companies. The regulation also contemplates that microgrids will be at least three-fourths powered by renewable resources, with non-renewable allowances for special modes of operation (e.g., during islanding), and for combined heat and power microgrids serving thermal loads. Table 3 outlines this approach to renewable capacity. Noting that PREPA's interconnection rules at the time specifically excluded microgrids, PREB ordered PREPA to prepare proposed microgrid-applicable interconnection standards within 120 days of Regulation 9028, and allowed microgrids to operate in islanded mode in the meanwhile [138]. The utility did not meet this deadline, and expressed concern about various financial, planning, rate, and operational issues associated with microgrids. After ordering PREPA to show cause for why it should not be fined for failing to prepare the interconnection standards, PREB took over the process of proposed rulemaking on interconnection [139].

Resilience in Puerto Rico's Microgrid Regulation
Regulation 9028 highlights the ability of microgrids to operate in island mode, provide services during grid outages, and to assist with more rapid restoration of service to consumers rather than waiting for reconnection to the grid at-large. However, as in Hawai'i, it does not appear that this resilience capacity will be monetized and compensated via a tariff. Rather, microgrids have been strongly encouraged via grid planning. After twice rejecting PREPA's integrated resource plan (IRP), PREB gave interim approval to a modified IRP and action plan with a specific directive to pursue microgrids: The Energy Bureau FINDS that microgrids form a critical part of the resiliency solutions envisioned for the Commonwealth. The Energy Bureau ORDERS PREPA to directly incorporate promotion of microgrid resources into all of its transmission, distribution, and resource planning exercises and all deployment actions taken in compliance with the modified Action Plan described by the Energy Bureau in this Final Resolution and Order. This includes facilitating timely and non-discriminatory access for all [distributed generation] and microgrid facilities to interconnect with PREPA's grid [140].
PREB further determined that "rapid deployment of points of distributed resiliency, including the use of microgrid, single-site solar PV and battery resources, or aggregated [virtual power plants] must form a part of PREPA's near-term approaches to developing a more resilient grid." This order rejected a prior proposal for a series of larger "minigrids," each served by natural gas-fired generation. PREPA was instead ordered to evaluate the development of one minigrid region, and to include microgrid deployment in that analysis.

Cost-Effectiveness in Puerto Rico's Microgrid Regulation
As with Hawai'i's tariff, Regulation 9028 does not contemplate specific compensation structures for microgrids that may supply energy or services to the grid at-large. Subsequently, however, Puerto Rico adopted the landmark Regulation on Wheeling in 2019 [141]. This regulation breathes life into Regulation 9028′s provision allowing microgrids to export energy and services to consumers other than microgrid participants and PREPA. The wheeling mechanism requires the unbundling of costs associated with generation, transmission, and distribution functions of the PREPA system. Microgrids and other independent power producers then must pay a wheeling charge (presumably passed on to wheeling customers) to compensate PREPA for the use of transmission or distribution infrastructure. Although it does not appear that the wheeling structure is yet complete, this dual development of regulatory frameworks for microgrids and wheeling puts Puerto Rico ahead of other U.S. jurisdictions in creating market mechanisms to promote microgrids that can export energy and services to the grid at-large. Ultimately, the cost-effectiveness of this approach will be resolved by the evolution of wheeling charges, and the market uptake of wheeled energy from microgrids. PREB has opined that the success of this model is dependent on properly allocating costs across customer classes, "to ensure that wheeling does not result in technical problems, rate increases, or any other unfair cross subsidization" [141,142]. O'Neill-Carrillo et al. noted that assessing the cost-effectiveness of renewable community microgrids in Puerto Rico's energy transition should also account for value streams capturing resilience, sustainability, and support for local socioeconomic activity [143].

Energy Justice in Puerto Rico's Microgrid Regulation
The destruction caused by Hurricanes Irma and Maria in 2017 illustrates that energy justice is intertwined with resilience. Román et al. found that some communities, e.g., those benefitting from distribution upgrades and backup generating capacity, achieved 60 percent recovery within days of Hurricane Maria [134]. They found that other communities, especially rural and low-income communities disadvantaged by density-based restoration protocols, did not have power restored for months. Despite this link, it does not appear that Regulation 9028 specifically targets or promotes microgrid development in these communities.
Some aspects of Puerto Rico's approach, however, may promote energy justice. Provisions allowing microgrids to wheel energy to retail customers are consistent with Wolsink's observation that microgrids can reorganize traditional roles in electricity production and consumption [86]. PREB's rejection of PREPA's proposed fossil fuel-based minigrid system may also reflect progress on energy justice. PREB acknowledged that "citizens were profoundly concerned about the IRP not considering the health risks associated with the construction of new fossil fuel generation infrastructure near populated areas" [140]. Advocates in the IRP process buttressed this concern by referencing a health study of various communities, which they asserted documented an increased prevalence of cardiovascular, respiratory, and other disease in the community of Guayama, downwind of a coal-fired power plant and other industrial facilities [144]. These advocates also noted that environmental justice communities sited near fossil fuel-fired power plants in Puerto Rico have mean household incomes far below average. Also, although such communities are entitled to access to information under the U.S. Emergency Planning and Community Right-to-Know Act, these advocates asserted that PREPA "historically has not complied with this requirement." This advocacy on behalf of environmental justice communities was accompanied by expert testimony on the value of distributed resources such as microgrids. PREB ultimately agreed, finding that "intervenor testimony compellingly demonstrates the inherent value of small-scale distributed resources in the form of microgrids, single-site solar PV and battery storage, and aggregated solar PV and battery storage (or [virtual power plants]) for Puerto Rico as a critical part of an overall solution to ensure resiliency." In this sense, it appears that PREB's order aligned with energy justice-oriented positions adopted in the proceeding. It remains undetermined whether the DERs-based vision set forth in the IRP order will be supplemented by ancillary solutions, such as access to capital for microgrid development in underserved communities, and a fully realized wheeling program. With respect to financing, advocates in the IRP docket pointed to then-proposed legislation that would have required PREPA to provide capital for the installation of distributed solar systems, and called on PREB to open a docket specifically designed to examine financing options [144]. It does not appear that such a process has commenced.

Conclusions
Renewable microgrids are emerging. Rising regulatory and development interest in places like Hawai'i, Puerto Rico, California, Korea (and other regions not otherwise identified in this paper, such as Vietnam and Australia) suggests that they will play a role in decarbonizing to satisfy 100% renewable energy standards. It appears that the potential resilience value of microgrids remains a primary motivator. That resilience motivation is not likely to retreat, particularly in a world where the COVID-19 pandemic is illustrating that reliable access to power is not a luxury, but rather is a necessity for access to schools, workplaces, and other human rights.
Despite this reality, policy remains clouded on how, or whether, the resilience value of distributed and networked renewable microgrids will accrue to the grid at-large, or will instead accrue primarily to microgrid participants. This uncertainty appears to be grounded in familiar territory. The experience in Hawai'i and Puerto Rico highlights unresolved fundamental tensions about using utility infrastructure to deliver non-utility energy and services. These tensions are grounded in energy justice considerations; they implicate questions about (i) how benefits and burdens associated with modern energy infrastructure will be distributed, and (ii) whether legacy utility models and infrastructure will evolve at the pace of rapid decarbonization and play a role (accelerative or otherwise) in efforts to restore front-line communities. Other energy justice issues also remain unresolved, such as how to ensure equitable access to capital for developing community microgrids.
Although these justice considerations may seem divorced from the types of technical analyses also reviewed in this paper, future research inquiries can play an important role in helping regulators and advocates resolve these tensions. Most immediately, there are gaps in the quantitative literature about evaluating the cost-effectiveness of renewable microgrids through a socio-techno-economic lens that captures a sufficiently broad set of costs and benefits relevant to microgrids (e.g., resilience, land use, sustainability, demand response capacity, frequency regulation, resource adequacy, spinning reserve, etc.). Until sufficiently holistic notions of costs and benefits are quantified, it seems likely that the potential of microgrids will remain uncertain, particularly for grid-connected microgrids in highly regulated jurisdictions.