Smaller, Cheaper, and More Resilient


The rationale for microgrids.

Fortnightly Magazine - April 2013

Hurricane Sandy’s impact in New York City, Long Island, and New Jersey revealed what industry experts have long known: utilities and their regulators have underinvested in the nation’s electricity infrastructure, creating a grid that is more unreliable than a modern economy needs. Sandy’s aftermath—with long delays in restoring power, communication lines out, world-class hospitals evacuating patients, sewage treatment plants dumping waste into waterways—underscored that the grid’s one-size-fits-all standard of reliability lacks justification in a world where some services are more important than others. The primacy of emergency services aside, many believe that we also need to re-evaluate how society prioritizes investment of rate-payer dollars to operate, restore, plan, and build the transmission and distribution components of the electric system.

While the grid has suffered from underinvestment, the technology sector has undergone revolutions in cellular communications, microprocessor capability, and Internet connectivity. These revolutions have enabled the development of microgrids—small-scale electricity systems for one or more large users, which combine efficient generation of power with its carefully monitored use, with demand response (DR) and energy efficient technologies, in a single geographic location. These microgrids operate in parallel with or islanded from the larger grid.

More importantly, microgrids can provide what the public now seeks uninterrupted service for critical institutions and key components of infrastructure even when power is lost to the larger grid. They also bring competition’s creative forces to the electricity grid, which has long suffered a dearth of innovation.

In a region rich in microgrids, critical institutions will remain operational during blackouts, renewables will be deployed with ease, emissions will decline per unit of power used, and the grid and institutions that depend on it will each become more efficient and more resilient.

Microgrids and Macro Benefits

Microgrids enable large public or private institutions—hospitals, water and sewer treatment plants, universities, economic centers, shelters, and housing complexes—to obtain a secure supply of power or to restore it more quickly in the event of a blackout. When the larger grid loses power, institutions with microgrids can remain operational for weeks. Microgrids also enable host institutions to manage their own electricity use and reduce costs. Microgrids also can benefit the overall utility grid by providing increased efficiency and reducing operational and capital expenses for some sites that are costlier to serve than the average customer site. And microgrids can benefit the public, by improving the ability to maintain critical operations.

Figure 1 - Microgrid Capacity by Region

A growing number of world-class microgrids exist today in the United States and across the globe. Universities, hospitals, and other institutions with a public service mission are increasingly deploying microgrids. The potential of microgrids was illustrated during the 2011 earthquake and tsunami in Japan, when amidst extraordinary devastation, the Sendai 1-MW microgrid at Tohoku Fukushi University operated for two days in islanded mode while the surrounding region was without power.

Figures 1 and 2 portray the regional and host institution characteristics of microgrids worldwide. The majority are found in North America; the plurality are found on campuses.

Microgrids will increasingly be deployed to capture the benefits of solar energy, particularly in those regions where the cost of generation from photovoltaics (PV) is approaching grid parity. The networked energy management and storage capability of microgrids enable them to take full advantage of this generation technology.

Operations, Revenues, and Siting

Institutions and other customers deploy microgrids to pursue various goals—from improving resilience and reliability, to improving the cost or environmental characteristics of their energy supply. As such, microgrids differ from one host institution to the next. Generally, however, a typical microgrid installation includes: on-site generation, often a cogeneration (also called “combined heat and power,” or CHP) plant that supplies both electricity and heat to the host institution, and thus is likely to be more efficient than the energy delivered by the current commercial grid; a network of sensors that monitor energy use and communicate with a central controlling device; advanced building systems that manage a variety of functions, such as lighting, traditional HVAC, and sometimes storage of cooling and heating capacity, all of which are centrally controlled; batteries and renewable energy technologies; and other devices and behavioral initiatives the host undertakes to increase efficiency. A microgrid’s electrical relationship to the larger grid is controlled by interface technologies at the point of interconnection.

Figure 2 - Microgrid Capacity by Market Segment

Microgrid revenues can be derived from some or all of: a) reductions in use, stemming from energy efficiency or DR initiatives; b) more efficient purchasing of electricity, including buying off-peak electricity to reduce peak demand, often by storing it with some form of storage of thermal energy—i.e., pre-heating or pre-cooling building spaces, or storing hot or chilled water; c) sales of electricity to the host institution; d) balancing generation and usage with control technology, and selling excess electricity to the grid; and e) capacity charges and other revenues stemming from the benefits supplied to the larger grid.

Microgrids create significant benefits when deployed in areas of the grid that need additional electrical support, or where they eliminate the need for costly grid upgrades—the so-called “non-transmission alternative” or NTA (see “Looking Beyond Transmission”). Many of these benefits accrue to the grid, not the microgrid host institution, and thus aren’t typically monetized for the benefit of the microgrid owner. To capture their full potential value, microgrids should be deployed in electrically appropriate areas of the larger grid, as was done recently by Central Maine Power and is somewhat more common in Europe.

Key choices faced by the microgrid host institution include: how much electricity to generate on-site; whether to operate islanded from or in parallel with the larger grid, and, if islanded, how frequently to operate in islanded mode, and toward what goals; whether to sell electricity, grid support services, or capacity to the larger grid, or to simply provide generation to meet the host institution’s needs; and whether to sell power to off-site users—that is, users not part of the host institution.

Microgrids operate in an intricate legal and regulatory environment—and arguably they’re proliferating in spite of it. State law determines whether a microgrid can sell electricity or steam to nearby institutions, or even branches of the host institution that happen to have another electric meter, or are located on the other side of a public street. State law and the local grid operator, which often is vested with quasi-legal authority owing to its position as a regulated monopoly, determine the extent to which a microgrid operator can sell electricity and related services to the larger grid and, in some cases, whether fully islanded operations are allowed. Federal law, as embodied in the regulations governing independent system operators and regional transmission organizations, also can inhibit microgrids when they require elaborate and expensive interconnection procedures. And microgrid operators must master a host of other legal and regulatory issues to build, operate, and maximize the benefits of their systems.

Sophisticated microgrids of all sizes are now operational throughout the United States, deployed by different host institutions with varying energy, financial, and operational goals. The multiple types of microgrids demonstrate the adaptability and flexibility of the model.

Islanding Universities

One of the most celebrated microgrids is found on the campus of the University of California-San Diego. The microgrid covers 1,200 acres and serves a daytime peak population of 45,000, plus laboratories that require an uninterruptible supply of power. The UCSD microgrid self-generates about 85 percent of its power, is capable of operating in island mode, and has a control technology that regulates energy use across the campus. UCSD claims it has helped it save almost $800,000 in monthly purchases of electricity. UCSD campus also has 11 percent of its power supplied by solar PV panels. It’s among the largest, most comprehensively designed, and renewable-enabling microgrids in the world.

Cornell University designed its microgrid after the blackout of 2003, when it experienced significant delays in the restoration of electricity. The university decided it wanted a secure supply of power at all times, and created a fully functioning microgrid. Today Cornell operates one of the largest microgrids in the state of New York. With several CHP generators on site, Cornell’s microgrid supplies electricity and heat to more than 150 buildings on campus, and cooling to another 75 buildings. The microgrid has optional islanding capability, but usually operates in parallel with the larger grid to supply all the campus’ electricity and thermal needs. This 35-MW system also has a control technology that regulates energy use across the campus.

The Illinois Institute of Technology is in the process of developing a prototype microgrid system—i.e., several microgrids linked together in a loop to provide a completely reliable source of electric power.1 The system is designed to “eliminate costly outages, minimize power disturbances, moderate an ever-growing demand, and curb greenhouse gas emissions.” It’s projected to cost $12 million and will pay for itself in five years. IIT’s confidence in the project’s economics suggests that financing a microgrid lies within the reach of many large, financially sophisticated institutions.

Absolute Security for Defense

The U.S. Department of Defense has embarked on a major initiative to deploy microgrids in bases across the United States. Its reasoning emphasizes the appeal of distributed energy systems for supply security. The DoD has determined that its “installations are largely dependent on a commercial power grid that is vulnerable to disruption due to aging infrastructure, weather-related events and a potential kinetic or cyber attack.” 2 Deployment of microgrids is the preferred strategy because “[o]n-site energy is critical to making our bases more energy secure.” Microgrids will ease integration of renewable generation, especially solar PV panels, building control systems, and advanced storage and efficiency technologies. Finally, an MIT study for the DoD concluded that “the most cost-effective microgrid solutions will be those that take into account the needs of the local commercial electric grid and implement their systems so that they can earn value helping to meet those needs.”3

The flagship DoD microgrid is the National Interagency Biodefense Campus at Fort Detrick in Maryland. On-site cogeneration plants provide electric power, steam, and chilled water to the base’s medical and research labs. Generation capacity is being expanded substantially from 7 MW to 17 MW. Owing to its large critical-mission load and the resultant necessity of a secure supply of power, the base’s microgrid supplies power with a 99.999-percent reliability factor. This extraordinary degree of energy security is costly: prevailing local electricity rates are approximately $0.08 per kWh; Fort Detrick spends over $0.21 per kWh. In this regard, the base is an outlier among DoD installations; owing to its unique mission, it ignores the DoD cost-reduction imperative.

A more typical microgrid installation is the Naval Support Facility Dahlgren, in Virginia, just south of Washington, D.C. This facility has created a network of controls to monitor use of electricity, natural gas, and hot water throughout its extensive campus. Its various energy systems have been linked in a cyber-secure manner. Finally, the original diesel generators, installed because of persistent grid reliability problems, are beginning to be linked together into a true microgrid, with the same control system used in the larger commercial grid. The Dahlgren microgrid operates in parallel with the utility system to reduce demand on the larger grid, and can operate in island mode during outages. While the Dahlgren microgrid doesn’t sell electricity or grid support services to the larger grid, it does offer huge DR capacity—14 MW—to the local utility. Revenue from this program pays most of the costs of the microgrid. Dahlgren’s base commander is considering whether to incorporate renewable power into the microgrid.

Hospitals and CHP

Hospitals are among the most important institutions in a civic emergency and are well suited to microgrids. Microgrids can serve a hospital as well as a small or large network of surrounding facilities that are affiliated—or not—with the hospital. Of course, the legal and regulatory challenges to a microgrid increase when a microgrid seeks to sell power to adjacent, unaffiliated customers, because the local utility might have the exclusive right to sell electricity to third parties. But this concern need not prevent the development of microgrids.

Utica College and Faxton-St. Luke’s Healthcare created a small, shared microgrid serving the energy needs of both institutions. The 3.4-MW system, operational since 2009, uses four small CHP generators to supply Utica College, the local hospital, and an affiliated nursing home. The microgrid provides more than 80 percent of the hospital’s energy needs, 75 percent of the college’s power, and 50 percent of the nursing home’s energy needs. The microgrid can operate islanded, if the grid loses power, but typically operates parallel to the grid, and when circumstance permit, it sells excess power back to the grid.

In another example, a large cogeneration facility, the Medical Area Total Energy Plant (MATEP), serves major hospitals affiliated with Harvard University, as well as of other facilities in the Longwood area of Boston. This district energy system serves the energy needs of several host institutions and surrounding buildings. MATEP provides steam, chilled water, and electricity to more than 9 million square feet of space in facilities in Longwood, and seeks to provide a highly reliable source of power at affordable prices to its customers. MATEP serves five major hospitals that have over 2,000 beds, as well as biomedical and pharmaceutical research centers. A little over two years ago, the entity that owns and operates MATEP was acquired by a joint venture between Morgan Stanley Infrastructure Partners and Veolia Energy North America.

Rebuilding New York’s Grid

The devastation caused by Superstorm Sandy revealed the need for a new approach to delivering utility services in New York. Mayor Michael Bloomberg, in a speech on Dec. 6, 2012, pledged to “modernize our energy infrastructure by incentivizing large buildings and hospitals to invest in co-generation systems… We will work with Governor Cuomo to explore how we can accelerate investments in distributed energy, microgrids, energy storage, and smart grid technologies.”4

Prospective deployment of microgrids in Long Island and New York City will have two principal goals: making the electric grid more resilient and, in case power is lost, making certain that key institutions and components of critical infrastructure retain their power and can respond to emergencies. Cooperation of the local utilities and large host institutions is vital to achieve these goals.

On Long Island, deployment will depend on critical choices by LIPA, National Grid, and potential host institutions. The electrical characteristics of Long Island’s grid enable full deployment of microgrids, including the sale of energy, capacity, and other grid-support services back to the bulk power grid itself. LIPA and National Grid should designate locations where they expect to recommend line upgrades requiring substantial capital costs. They also should identify locations where, for technical reasons, microgrids’ substantial reductions in load and other efficiencies would benefit the grid. These locations are prime targets for microgrids. Potential host institutions should evaluate whether a microgrid is appropriate.

Host institutions that might be well suited for a microgrid include: the Brookhaven National Laboratory; Hofstra University; the SUNY Stony Brook Campus; Nassau Community College; North Shore LIJ Medical Center; the Stony Brook Medical Center; South Nassau Community Hospital; Good Samaritan Community Hospital; Winthrop Community Hospital; and Hauppauge Industrial Park, as well as other institutions of similar size.

The best locations are those where the larger grid can benefit from a microgrid and where a significant local institution is committed to serving as a host. Also, deployment of microgrids can expand on the distributed energy assets that already exist. For example, the 45 MW of generation at SUNY Stony Brook can be upgraded to more effectively serve the campus, the surrounding area, and become a model for the deployment of microgrids across the SUNY system.

There also might be a federal role. Given Sandy’s effect on Long Island, the U.S. Department of Energy and the Department of Homeland Security might be able to provide technical support, funding, or both to speed the local deployment of microgrids. The DOE’s Brookhaven National Laboratory could serve an ideal location for a DOE flagship microgrid.

Likewise, in New York City, deployment will depend on critical choices by Con Ed and potential host institutions. However, the electrical characteristics of Con Ed’s underground distribution system can make the sale of electricity and support services back to the grid technically more challenging than usual. The enthusiastic cooperation of Con Ed will be vital to ensure that microgrids can be deployed to the fullest extent possible. Even if that weren’t to occur, very large microgrids of 25 MW and higher likely could be deployed without undue technical limitations, and smaller microgrids could deploy with all but commodity and grid support, so-called “ancillary service” sales. Potential hosts include:

  • Private universities, like NYU5 and Columbia, each of which already has substantial on-site generation assets that could transition to a full microgrid, as well as others, including Fordham, St. John’s, Yeshiva, and Pace; 
  • Hospitals, whether affiliated with universities, operated as a group like Continuum, or part of the NYC Health and Hospitals Corp. A joint microgrid including Bellevue Hospital, NYU Langone Medical Center, and the large NYC men’s shelter that lies directly between them, could be ideal;
  • Housing complexes, from large NYCHA developments to private stand-alone complexes, like Stuyvesant Town, Peter Cooper Village, or Battery Park City;
  • Public universities, such as many of the CUNY campuses;
  • Economic centers like the New York Stock Exchange, data and communication centers, Rockefeller Center, the Brooklyn Navy Yard, and Hunts Point Distribution Center; and 
  • Sites with critical infrastructure, such as water and sewage treatment plants, shelters and schools, and police and fire command centers.

Whether any microgrid projects will emerge from the Bloomberg Administration’s current review remains uncertain. Also, since Bloomberg is now in his final term as mayor, the danger exists that current initiatives won’t survive into the next administration. It will be important to examine the findings of New York City’s infrastructure effort and incorporate the significant recommendations into any long-term plan developed by Gov. Cuomo’s administration. Additionally, the services of the New York Power Authority and NYSERDA will be important in evaluating which microgrids, if any, should receive public support.

Technology’s Unstoppable Force

Three barriers have prevented the wider deployment of microgrids.

First, potential customers face significant barriers to entry—including difficulties in learning about the microgrid product. Baldly stated, information about microgrids’ costs and benefits remains hard to obtain and difficult to understand. The host institution typically isn’t in the electricity business, and so it requires an unusual degree of leadership for a major health care, educational, or housing provider to understand microgrids’ benefits and then to make the decisions necessary to secure them. It’s no accident that many microgrids are deployed by universities with significant internal expertise in engineering and the hard sciences.

Second, pricing isn’t transparent. Reliability doesn’t have a price under the current grid’s regulatory regime. It’s accordingly difficult to price increased reliability. Moreover, improved reliability has broad social benefits—such as a hospital that functions during an emergency—which the host institution can’t identify or monetize.

Third, regulations constrain microgrids’ full development. Even if customers can learn about microgrids and decide that the benefits of a microgrid exceed its costs, the optimal deployment of a microgrid requires both an understanding of bulk power pricing—a notoriously non-transparent market—and an ability to work with local utilities and with the ISO or RTO, many of whom might view microgrids as reducing sales of electricity or damaging to wholesale market design. In such conditions, microgrids are difficult to deploy to their greatest advantage, undercutting their appeal.

Given the somewhat dysfunctional electricity markets, government’s role is clear. Removing informational barriers and pricing public goods are part of government’s mission. Government also should enact legislation and promulgate regulations to incentivize regulated monopolies to act in the public interest. Specifically, if microgrid expansion enhances electric security and thereby economic security, that expansion should be encouraged, and not discouraged as is likely under existing regulatory regimes.

A phased approach might prove to be the most successful, with government first leading the installation of microgrids in public colleges and universities as well as hospitals and other components of critical infrastructure. As these changes take root, they will create a path for more long-term legislative and regulatory reforms. It will be essential to develop partnerships with utilities from the beginning.

Finally, microgrids face challenges because they’re unique components of electricity infrastructure. They don’t fit within the governing central station model of power generation: electricity generated by large plants, sent long distances over transmission lines, and finally fed through distribution lines to passive end users. As such they don’t fit easily within the traditional utility operating model, and nor do they fit within the regulatory construct that has grown up around this model.

The conventional view is well-known within the industry. Incumbent utilities might find microgrids somewhat antagonistic to their business model. Microgrids provide generation capacity, but not within the central station model; microgrid assets might be owned by important institutions that are neither utilities nor connected to the industry; and microgrids arise from a free-market economic interest and technological capability to drive down energy use and increase the efficiency and resilience of energy services.

But this conventional view is dated. Technology here—as elsewhere—intrudes with unstoppable force.

Utilities that accept and embrace microgrids as part of 21st-century energy infrastructure will have a major role in their deployment, for several reasons. First, deployment of microgrids is technically complex. Utilities should be involved in siting them—as they are in siting new generation via the procedures of the ISO-RTOs—to ensure they benefit the larger grid. Second, utilities should help set standards for their design and site-specific configuration so islanding happens quickly and smoothly, and backup generators don’t fail as they did at the NYU Langone and Bellevue hospitals during Superstorm Sandy, forcing patients to be evacuated. Third, utilities could share in the financing of microgrids with host institutions, provided that the public benefits of the particular microgrid deployment are clear and quantifiable.

However, a few caveats are in order. First, utilities aren’t likely to be allowed to monopolize the microgrid business. Much like federal law prevents common control of transmission and generation assets within an organized wholesale power market to ensure healthy competition, many state laws will prohibit monopoly utility ownership of microgrids. Second, while the view that microgrids threaten utility sales is accurate, other factors—such as the deployment of power-hungry data centers—will help maintain underlying levels of electric demand, especially in urban areas.

One of the benefits of microgrids is their extraordinary ability to create efficiencies in and reduce the costs of electricity usage. By allowing utilities to invest in microgrids—without giving them a monopoly—regulators can create a new regime that fairly compensates utilities and their investors for participating in this new world.


1. See

2. Statement of Dr. Dorothy Robyn, Deputy Under Secretary of Defense (Installations and Environment) before The House Armed Services Committee, Subcommittee on Readiness, March 29, 2012 at 1.

3. Microgrid Study: Energy Security for DoD Installations, Lincoln Laboratory, Massachusetts Institute of Technology, June 18, 2012 at iv.

4. See New York City Government press release, PR 459-12.

5. NYU reportedly operated its cogeneration system in island mode during Superstorm Sandy, and maintained service to the university’s main campus at Washington Square and some other large buildings.