1,074 Microgrids: An emerging paradigm for meeting building electricity and heat requirements efficiently and with appro-

ppro- Marnay, Chris

priate energy quality Firestone, Ryan
1,074 Marnay, Firestone

Microgrids: An emerging paradigm for

meeting building electricity and heat
requirements efficiently and with appropriate
energy quality
Chris Marnay
Berkeley Lab MS 90R4000
U.S.A.
C_Marnay@lbl.gov

Ryan Firestone
Berkeley Lab MS 90R4000
U.S.A.
RMFirestone@lbl.gov

Keywords
combined cooling, heating and power, commercial buildings, not simply assume gas arrives at the gas meter, electricity at its
community energy systems, microgrids, engine generators, meter, and the two systems are virtually independent of one
gensets, on-site generation, photovoltaic, power quality and another; rather, energy conversion, heat recovery and use, and
reliability renewable energy harvesting may all be taking place simultane-
ously within the building energy system; 2. the structure of en-
ergy flows in the building must accommodate multiple energy
Abstract processes in a manner that permits high overall efficiency; and
The first major paradigm shift in electricity generation, deliv- 3. multiple qualities of electricity may be supplied to various
ery, and control is emerging in the developed world, notably building functions.
Europe, North America, and Japan. This shift will move elec-
tricity supply away from the highly centralised universal serv-
ice quality model with which we are familiar today towards a Introduction
more dispersed system with heterogeneous qualities of service. This paper examines the role of the μGrid (microgrid) para-
One element of dispersed control is the clustering of sources digm in revolutionising the current universal centralised model
and sinks into semi-autonomous μgrids (microgrids). Re- of electricity generation and delivery. Then it provides a survey
search, development, demonstration, and deployment (RD3) of several international research projects that pioneer research,
of μgrids are advancing rapidly on at least three continents, development, demonstration, and deployment (RD3) of μgrid
and significant demonstrations are currently in progress. This concepts as an alternative approach to integrating small-scale
paradigm shift will result in more electricity generation close (< 1 MW) distributed energy resources (DER) into commercial
to end-uses, often involving combined heat and power applica- buildings with peak electrical loads of less than about 2 MW
tion for building heating and cooling, increased local integra- or into multi-family buildings or housing estates. A μgrid is a
tion of renewables, and the possible provision of heterogene- grouping of generating sources and loads operating semi-inde-
ous qualities of electrical service to match the requirements of pendently of the legacy power system, or macrogrid, typically
various end-uses. In Europe, μgrid RD3 is entering its third interconnected at a single point of common coupling (PCC).
major round under the 7th European Commission Framework The μgrid may include traditional reciprocating engine gen-
Programme; in the U.S., one specific μgrid concept is undergo- erators (gensets), microturbines, fuel cells, photovoltaic mod-
ing rigorous laboratory testing, and in Japan, where the most ules (PV) or other small-scale renewables, heat recovery from
activity exists, four major publicly sponsored and two privately thermal generation and use, electrical and heat storage devices,
sponsored demonstrations are in progress. This evolution poses and controllable end-use loads. Three likely μgrid features are:
new challenges to the way buildings are designed, built, and 1. efficiently meeting total system energy requirements, often
operated. Traditional building energy supply systems will be- by including combined heat and power (CHP) technology,
come much more complex in at least three ways: 1. one can- especially for building heating and/or cooling, 2. providing

ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT! 31

1,074 MARNAY, FIRESTONE PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY

heterogeneous levels of electricity security, quality, reliability times bring our ability to maintain current SQRA standards
and availability (SQRA) that match the requirements of vari- into doubt.
ous end-uses, thereby potentially lowering expectations for
improvements in the macrogrid to meet the needs of a digital
society, and 3. appearing to the macrogrid as a controlled entity,
Two Visions of the Future Grid
akin to a current local utility customer, or conversely akin to Two alternative visions in current currency of how the power
a small embedded generation source, if the μgrid exports. The system might be retooled to provide high SQRA are a super-
materials presented are based on presentations at a series of grids view and a dispersed paradigm. These are obviously only
international symposiums held in the U.S. in 2005, in Canada two of many possible paths and full justice cannot be given here
in 2006, and a third in Nagoya, Japan, in April 2007. Materials to the technical intricacies of any specific vision. The intent is
from these events can be found at http://der.lbl.gov only to contrast in a comprehensible way the central theme
of two divergent alternatives. For more detail on a supergrids
view, see Gellings et al (2004), Amin (2005), or Amin and Wol-
Dispersed Generation Paradigm Shift lenberg (2005). A comprehensible vision for a dispersed grid is
Trends emerging in the power system suggest that the highly presented by the European Commission (2006), or, for other
centralized paradigm that has dominated power systems for the voices from the dispersed camp, see Lasseter (2006) or Marnay
last century may eventually be replaced, or at least diluted, by and Venkataramanan (2006), but these are by no means the
an alternative. In the new paradigm, control is more dispersed, only contributors to this ongoing debate.
and universal SQRA is replaced by heterogeneous service tai-
lored to the requirements of highly diverse classes of end-uses. SUPERGRIDS VISION
This shift may be thought of as comparable to the replacement A supergrids vision is shown in Figure 1. The x-axis of Figure 1
of centralised computing by desk and laptop computers, or the shows the history of the current centralised paradigm, and the
switch from land based telecommunications to mobile devices. y-axis reliability expresses as nines, e.g. 3 nines implies 99.9 %
Our current power delivery paradigm has been in place world- availability. The equivalent annual expected outage times are
wide for a long time, i.e. since the emergence of polyphase AC shown for reference. The SQRA of delivered electricity has mul-
systems around the turn of the last century. SQRA targets are tiple dimensions, e.g. voltage swells and sags, harmonic distor-
consistent virtually all across vast regions, e.g. all of North tion, etc. Reliability is used here as a representative dimension
America, and where standards cannot be met, it is usually the because it is much more easily comprehended than others,
result of a local technical difficulty and not the outcome of a and we have some intuitive sense of its historical trajectory, as
deliberate attempt to deviate from the norm. Emerging changes shown. In the early days of centralized power systems, electric-
on the demand-side include our seemingly unquenchable thirst ity was supplied by small local stations with only very few gen-
for electricity, in large part driven by the increasingly dominant erators to a limited number of customers, in the very early days
role of commercial building use in post-industrial economies, using DC. These highly unreliable systems were consolidated
by an emerging digital age that is significantly tightening our by ones covering large areas based on Nikola Tesla’s concepts
SQRA requirements, by the emergence of viable small-scale for large-scale AC systems. This interconnection naturally im-
fossil generation often with power electronics and CHP, and proved reliability because many more generators were simul-
by an urgent need to incorporate small-scale renewable gen- taneously available. The green arrow shows how this process,
eration to abate carbon emissions. Meanwhile, on the supply- together with significant and steady technological progress,
side, concerns about terrorism, restrictions on system expan- resulted in steadily improving reliability, reaching the levels
sion, and the uncertainties of volatile markets in energy-short experienced in North America today; however, note that reli-

Figure 1. Schematic of a Supergrids Vision

32 ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT!

PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY 1,074 MARNAY, FIRESTONE

Figure 2. Schematic of a Dispersed Vision

ability is considerably better in western Europe, and better still deployed to improve on the existing system’s weakest link; and
in some Asian countries, notably Japan. The red arc reflects the second, widespread use of supply and other resources close to
great concern in North America following the huge New York sensitive loads protect them at the levels they demand. This
blackouts of the 1970s that backup gensets or other emergency is shown in the figure as local heterogeneous SQRA. In other
sources be provided to critical loads, and such requirements words, end-uses are serviced with SQRA tailored to their re-
became embedded in building codes. Thus, over the last quar- quirements. In a sense, this vision is one of increasingly het-
ter century or so, a separate higher reliability service has been erogeneous SQRA downstream in the power system. The tra-
introduced by installing generation close to sensitive loads. ditional universal SQRA is retained in the high voltage meshed
The supergrids camp holds that deployment of diverse suites grid, but the distribution network has differing levels of invest-
of new technologies can significantly improve the performance ment in equipment to enhance SQRA, and finally within cus-
of all elements of power systems built around the traditional tomer sites, SQRA is ultimately matched to end-uses by means
paradigm; i.e., delivered SQRA can be dramatically improved of segregated circuits or by provision of high quality service at
within the existing framework. In the schematic, this is shown the point of end-use, either by microgrids or power condition-
by the arcing curves into the future. While much of the im- ing equipment. In this dispersed paradigm, μgrids enter in two
provement inevitably must come in the distribution system ways, as coordinated groupings within the distribution net-
because most outages and power quality problems occur there, work that can operate semi-autonomously of the high voltage
over 90 % of interruptions in the case of North America. Distri- meshed grid upstream of substations, and downstream of the
bution represents the most vulnerable link in the delivery chain meter where sources and sinks are organized to jointly provide
because of its sheer size and dispersion, as well as its exposure heat and electrical energy, as well as heterogeneous SQRA.
to the myriad hazards of extreme weather, accidents, and mis-
chief. Even in the supergrids view, inevitably there will be end-
uses that require SQRA beyond even the performance of the
Europe
much enhanced delivery chain, but these can be kept to a mini- Early μgrid RD3 in Europe occurred within the 5th Framework
mum; i.e., the gap between dashed curves can be kept small. Programme (1998-2002). A Consortium led by the National
This vision imagines massive investments in new technologies Technical University of Athens (NTUA) included 14 part-
for electricity delivery, such as superconducting lines, etc. ners from 7 EU countries, including utilities, e.g. Électricité
de France, equipment manufacturers, e.g the German power
DISPERSED GRID VISION electronics company SMA, and research institutions and uni-
In this view, traditional universal service upstream of the sub- versities, e.g. Labein. The main objectives were to study high
station is not improved significantly but rather holds steady renewable and other microsource penetration into the grid,
at current levels, as seen in Figure 2 as universal homogeneous μgrid islanding operation, and μgrid controls. Several levels
SQRA. In other words, operation of the high voltage transmis- of centralized and decentralized control were explored at sev-
sion system and everything upstream of it are operated as now, eral laboratories, notably the Institut für Solare Energiever-
with similar rules and conventions, and similar SQRA stan- sorgungs-technik at the University of Kassel, the University of
dards. Sensitive loads are then increasingly served locally in Manchester, and the National Technical University of Athens.
two ways: first, improvements in the distribution system are A follow up project was completed within the 6th Framework
Programme (2002-2006), again led by NTUA but with a some-

ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT! 33

1,074 MARNAY, FIRESTONE PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY

Figure 3. The Kythnos Microgrid

Figure 4. Mannheim-Wallstadt Microgrid

what different, although diverse, group of partners. This effort tribution transformer through four feeders, each about 400 m.
focused on new micro-sources, storage, and control. There was Using a power electronic flexible AC distribution system and
also considerable effort on network design, protocols, and the storage, islanded operation of the μgrid and power quality con-
benefits and costs of μgrids (Hatziargyriou 2006). A new round trol will soon be tested. A third project in Germany, shown in
of projects will soon begin under the 7th Framework. Figure 4, at the 400-inhabitant Am Steinweg residential estate
Several pilot μgrid installations have been completed. As has 69 kW of DER including a 28 kW CHP plant, 35 kW of PV,
shown in Figure 3, twelve houses in a small valley on Kythnos and an 880 Ah battery bank. Other projects include an ecologi-
Island in the Cyclades Archipelago of Greece are supplied by a cal estate in Mannheim-Wallstadt, as shown in Figure 4, and
μgrid composed of 10 kW of PV, a 53 kWh battery bank, and a projects in Denmark, Portugal, Spain, and Italy. In addition to
5 kW diesel genset. The μgrid includes 3 SMA3.6 kW inverters these EC projects, relevant European demonstrations are also
connected in parallel to form one strong single-phase circuit in being conducted at the national or local government levels.
a master slave configuration. The most innovative aspect of this
system is that the battery inverters operate in frequency droop
mode without fast electrical controls. This approach passively
North America
allows information flow to μgrid devices, in a manner similar The U.S. has a slowly expanding μgrid research program, sup-
to the Consortium for Electric Reliability Technology Solutions ported both by the U.S. Department of Energy under the Office
(CERTS) approach used in the U.S. demonstration, described of Electricity Delivery and Energy Reliability, and by the Cali-
below (Engler 2006/7). fornia Energy Commission through its Public Interest Energy
A second demonstration has been conducted at the Contin- Research Program. The most well known effort has been pur-
uon holiday camp in the Netherlands, which has more than 200 sued under the Consortium for Electric Reliability Solutions
cottages equipped with a total of 315 kW of PV modules, in- (CERTS, http://certs.lbl.gov ), which was established in 1999 to
terconnected by inverters. The cottages are connected to a dis- explore implications for power system reliability of emerging

34 ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT!

PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY 1,074 MARNAY, FIRESTONE

Figure 5 Schematic of the CERTS Microgrid

technological, economic, regulatory-institutional, and environ-

mental influences. From its inception, the likely emergence of
DER was recognized as an important factor affecting reliabil-
ity, and it has consistently been a feature of the CERTS RD3
portfolio.
The specific concept of the CERTS Microgrid (CM) shown
in Figure 5 was fully developed by 2002, and was described in a
white paper presented at a CEC workshop on 2 May 2002, after
which building physical examples was undertaken (Lasseter et
al 2003). As with most μgrid paradigms, the CM is intended
to seamlessly separate from normal utility service at a single
point of common coupling (PCC) during a disruption and con-
tinue to serve its critical internal loads until acceptable utility
service is restored. As in the Kythnos demonstration, the CM
provides this function for relatively small sites without need for
costly fast electrical controls or expensive site specific engineer-
ing. Unlike Kythnos, no single device is essential for operation,
creating a robust system. To the utility, the CM appears as a
single controlled load, and it is explicitly designed to provide
Figure 6. Layout of the CERTS Microgrid Test at the Dolan
heterogeneous SQRA as varying reliability on circuits. In the
Technology Centre
event of a grid disturbance, a static switch opens and the CM
serves critical loads in islanded operation until grid power is
again adequate, and reconnection occurs. This switching tech-
One notable feature of the CM project has been simultane-
nology is critical to the CM and many other μgrid concepts,
ous RD3 into necessary tools for μgrid deployment, other than
and has been the focus of intensive RD3 in the past few years
the actual electrical hardware. Two major products of this uni-
(Lynch, et al, 2006). Several technologies are available for fast
fied approach are the μGrid Analysis Tool, under development
switching, but at considerable cost. The CM is also a dispersed
at the Georgia Institute of Technology, and the Distributed En-
plug-and-play system, i.e. no custom engineering is required
ergy Resources Customer Adoption Model (DER-CAM) in use
for interconnection of any single device, as long as it has CM
at Berkeley Lab and several other R&D facilities worldwide.
capability, making system configuration flexible and variable.
DER-CAM is discussed further below. The CM is quite explic-
Sources may not only be spread across circuits, they may be
itly designed around CHP applications and hence analysis of
physically placed around the site, quite possibly co-located with
CHP applications in buildings is a central part of the CERTS
convenient heat sinks that offer economically attractive CHP
RD3 program.
opportunities. Finally, the CM has generic slow controls. The
In Canada, μgrid RD3 activities concentrate on the medium-
CM is currently undergoing testing at the Dolan Technology
voltage (25-170 kV) distribution network (Katiraei and Iravani
Center in Columbus, OH, using three reciprocating engine
2007). For example, the Fortis, Alberta, distribution system is
gensets as prime movers, as shown in Figure 6. If this test is
a grid-interface microgrid composed of a 25 kV distribution
successful, a full-scale field demonstration will follow.
network that is normally connected to the substation. One ap-

ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT! 35

1,074 MARNAY, FIRESTONE PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY

As shown in Figure 8, the first of Japan’s μgrid demonstra-

tion projects started during the 2005 World Exposition, us-
ing a combination of varied chemistry fuel cells, 270 kW and
300 kW Molten Carbonate Fuel Cells (MCFC), four 200 kW
Phosphoric Acid Fuel Cells, and a 25 kW Solid Oxide Fuel
Cell (SOFC). The MCFCs use a gas derived from wood waste
and plastic bottles. Experimental intentional islanding has also
been conducted. Recently, the system was permanently moved
to the Central Japan Airport City in Nagoya, where it will sup-
ply a Tokoname City office and a sewage treatment plant using
a private feeder.
The Hachinohe, Aomori Prefecture, project began operation
in October 2005 and is being evaluated for SQRA, cost effec-
tiveness, and carbon emissions reduction over its demonstra-
tion period stretching through March 2008. The μgrid has PV
Figure 7. Locations of the NEDO Microgrid Demonstration and wind turbines totalling 100 kW, 510 kW of controllable
Projects engine gensets supplied by digester gas from a sewage plant,
and a 100 kW lead-acid battery bank. Seven Hachinohe City
buildings are supplied via a private 6 kV, 5.4 km distribution
proach to maintain the supply during substation maintenance
feeder, with the whole system connected to the commercial
periods or subsequent to faults in the main grid is to tempo-
grid at a single point. Test islanding operation is also planned
rarily connect the distribution system to an alternative 25 kV
for this project.
feeder. Another approach to supply the load is to form an is-
In a third NEDO project, the municipal government of Kyo-
land on either the entire or a portion of the distribution feeder,
tango City, north of Kyoto, leads a virtual μgrid demonstration.
depending on adequate availability of power from local DER
The DER included are 50 kW each of PV and wind turbines,
units. At a notable example in British Columbia, the Hydro
five 80 kW biogas engines, a 250 kW MCFC, and 100 kW of
Boston Bar system uses a planned islanding option.
battery back-up. In this project, an energy centre communi-
cates with the DER over the existing utility network to coor-
Japan dinate demand and supply. Imbalances between supply and
Worldwide, μgrid RD3 is most active in Japan. To increase demand are resolved within five minutes.
potential renewable energy harvesting near demand centres, Finally, NEDO sponsors an ambitious and interesting multi-
Japan’s μgrid RD3 focuses on utilising controllable prime mov- ple SQRA demonstration project in Sendai, Miyagi Prefecture,
ers, such as natural or biogas fired gensets, to compensate for shown in Figure 9. This μgrid demonstrates multiple SQRA on
variable demand and local small-scale intermittent renewable adjoining rest home, high school, university, and waste treat-
supply. The New Energy and Industrial Technology Develop- ment facilities. The energy centre and a dedicated distribution
ment Organisation (NEDO), the research funding arm of the line are connected at a single PCC. The main DER are a 250 kW
Ministry of Economy, Trade and Industry, has started four MCFC, two 350 kW natural gas-fired gensets, 50 kW of PV,
demonstrations, as shown in Figure 7 (Funabashi and Yokoya- and batteries. These should be apparent in the figure. The lead
ma 2006). organization for this project is NTT Facilities, an arm of Japan’s
telecom giant. Because of this industry’s expertise in the high

Figure 8. The Aichi Microgrid Installed at the 2005 World Exposition

36 ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT!

PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY 1,074 MARNAY, FIRESTONE

Figure 9. The Sendai Multiple SQRA Microgrid Demonstration

SQRA DC systems that have always powered telephone service vary widely across grids. The product of these factors means
worldwide, the Sendai demonstration features direct service to the carbon footprint of commercial buildings can grow rapidly,
DC telecom loads. As well as DC, multiple qualities of stand- but changes in the fuel mix, e.g. more natural gas fired genera-
ard AC service are delivered from the clean power building tion, can also have a big effect. Further, in warm climates such
marked in the rear of the compound, creating an outstanding as most of the U.S. and Japan, and for an increasing share of
example of heterogeneous SQRA. In fact, this μgrid supplies Europe, commercial sector cooling is a key driver of peak load
AC to the nearby buildings at four different service qualities. growth, and hence the stress to and investment in the wider
A premium quality A service for critical loads is never inter- power system. Consequently, deployment of μgrids to serve
rupted, and waveform correction is performed on it. When the buildings, especially ones applying CHP technologies for cool-
utility grid has a momentary voltage sag or outage, the three ing, is central to containing the growth of electricity consump-
B quality circuits receive SQRA. The B service is further sub- tion and its associated carbon emissions. DER-CAM has been
divided into three different types. During outages, the higher developed as part of the CM RD3 project specifically to analyze
quality B1 service is backed up by storage, while B2 is backed the economics of building-scale μgrids.
up by distributed power, i.e. slower responding backup, while
B3 service is not backed up and experiences grid SQRA (Hirose
et al 2006). Note the similarity between this arrangement and
DER-CAM
the multiple SQRA on the various circuits of the CM DER-CAM identifies optimal technology-neutral μgrid invest-
In addition to the government-sponsored projects described ments and operating schedules at a given site, based on available
above, there are significant research activities in Japan’s private equipment options and their associated capital and O&M costs,
sector. Shimizu Corporation, a large construction company, is customer load profiles, energy tariff structures, and fuel prices.
developing a μgrid control system at its Tokyo test facility. Also, The Sankey (Spaghetti) diagram in Figure 10 shows partially
Tokyo Gas, together with the University of Tokyo, plans to es- disaggregated site end-uses on the right-hand side, and energy
tablish a μgrid to supply three-level power quality to a building inputs on the left. As an example, the refrigeration and cooling
of the Yokohama Research Institute. load may be met in one of multiple ways, including standard
electrically powered compressor cooling, direct fire or waste
heat activated cooling, or direct engine powered compressor
Building CHP cooling. DER-CAM solves this entire problem optimally and
While this paper has focused on μgrid demonstrations around systemically. Figure 11 shows a high level schematic of inputs
the world, and primarily on their SQRA aspects, equally im- to and outputs from the model.
portant from the buildings perspective is the CHP opportuni- As can be seen in Figure 11, DER-CAM picks its optimal
ties that μgrids will create. The importance of the commercial combination of μgrid equipment using hourly building end-
sector in electricity consumption in developed countries can be use loads and careful consideration of the detailed tariffs the
seen by three multiplicative factors. 1. The share of all energy building faces. The choice of technologies can include a broad
being consumed as electricity increases, e.g. in the U.S. from range of μgrid technologies. DER-CAM is particularly suited
13 % in 1980 to about 20 % today. 2. The commercial sector to evaluating μgrid CHP opportunities since it selects the op-
uses a growing share of all electricity, e.g. in the U.S. from 27 % timal combination of investment options, fully taking their
in 1990 to 35 % in 2005. And 3., typically an increasing share of interdependence into account; e.g., if there is a trade-off be-
electricity is generated thermally as carbon-free hydro sources tween thermally activated cooling and on-site genset capacity,
are fully exhausted, although the shares of carbon-free nuclear DER-CAM obtains the combination of the two that minimizes

ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT! 37

1,074 MARNAY, FIRESTONE PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY

Figure 10. Spaghetti Diagram of Energy Flows Through a Building

Figure 11. Schematic of DER-CAM Data Inputs and Outputs

cost. Thus, optimal combinations of equipment involving PV, cific Gas and Electric Company. This hypothetical facility has
thermal generation with heat recovery, solar thermal collec- 23 000 m2 of floor space and a peak electricity load of 690 kW.
tion, and thermally activated cooling can be identified in a way In this example, storage proved unattractive at current costs. To
that would be intractable by trial-and-error testing of all pos- exercise the model, both avoidable electrical and thermal stor-
sible combinations. However, DER-CAM currently has only age costs are set to zero plus an avoidable US$ 40/kWh cost.1
very limited capabilities for evaluating the SQRA benefits of The chosen optimal system consists of a natural gas-fired
μgrids. Such benefits can be considered only if a known cost, genset, solar thermal collectors, an absorption chiller and both
e.g., added equipment performance, can be directly traded off electrical and heat storage. Relative to a standard utility energy
against a known benefit. Work is currently under way to incor- supply-only case, the expected annual savings for the optimal
porate more SQRA capabilities. μgrid are $ 53 000/a (11.5 %), and the elemental carbon emis-
sions reduction is 59 t/a (10.4 %). Note that utility electricity
supply in San Francisco is relatively low carbon because of
Example DER-CAM Analysis the preponderance of natural gas as the marginal generation
Technology options in DER-CAM are categorized as either dis- fuel, which limits carbon savings from on-site generation. Fig-
cretely or continuously sized to reflect how closely to the opti- ures 12 and 13 show example DER-CAM operating results for
mal installed μgrids size is physically possible. This distinction the thermal and electrical balances of the hotel on typical days
is important to the economics of μgrids because equipment in January and July 2004. Note that the optimal technologies
typically becomes more expensive in small sizes. Discretely are a 200 kW reciprocating engine, a 585 kW (166 refrigera-
sized technologies are those which would be available to cus- tion tons) absorption chiller, 722 kW of solar thermal collec-
tomers only in a limited number of sizes, and DER-CAM must tors, 1100 kWh of electrical storage, and 299 kWh of thermal
choose an integer number of units, e.g., gensets. Continuously storage. While the economics of this case are not compelling,
sized technologies are available in such a large variety of sizes even with heavily subsidized storage, it is presented in detail to
that it can be assumed capacity close to the optimal could be demonstrate the scheduling capability of DER-CAM.
acquired, e.g. battery storage.
An example analysis was completed of a prototypical San
Francisco hotel operating in 2004, under the tariffs of the Pa-
1. At time of writing, 1 US$ = 75 EUR cents

38 ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT!

PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY 1,074 MARNAY, FIRESTONE

heat load heat and abs. chiller load non-cooling electric load total electric load
1200 1200
thermal power (kW)
1000 1000

electricity (kW)
thermal absoprtion cooling offset
800 storage 800
storage charging storage
600 natural gas 600 storage charging
combustion
400 400 generate
solar thermal
200 CHP heat 200 purchase
0 0
1 5 9hour 13 17 21 1 5 9hour 13 17 21

Figure 12. Meeting Thermal and Electricity Loads On a January Day

heat load heat and abs. chiller load non-cooling electric load total electric load
1200 1200
storage charging
thermal power (kW)

1000 1000

electricity (kW)
thermal storage charging
800 storage 800 absoprtion cooling offset storage
600 solar thermal 600 storage charging

400 400
200 CHP heat generate
200
purchase
0 0
1 5 9 hour13 17 21
1 5 9hour 13 17 21

Figure 13. Meeting Thermal and Electricity Loads On a July Day

The area underneath the solid black line in these figures is research results. Further, coordinated joint RD3 efforts among
the hourly energy demand. Area above the solid black line in- the major countries are emerging and are expected to provide
dicates storage charging. The various patterns in the graphs further mutual benefits in the historic effort to achieve the big-
indicate the source of the energy. For electrical loads the lower gest paradigm shift in electricity generation and delivery in a
profile indicates the portion of the electric load that can be met century or so, one that can accelerate lowering the carbon foot-
by only electricity, whereas the solid line above it is the total print of electricity supply and simultaneously meet developed
electric load, including cooling. Note that electric cooling loads countries’ growing requirements for high SQRA service.
can be offset by the absorption chiller. For thermal loads, the This power supply evolution poses new challenges to the way
lower line indicates the heat required for heating, whereas the buildings are designed, built, and operated. Traditional build-
solid black line indicates the total thermal load, including heat ing energy supply systems will become much more complex
required for the absorption chiller. in at least three ways. First, architects and building engineers
cannot assume that as now gas will arrive at the gas meter, elec-
tricity at its meter, and within the structure, the two systems are
Conclusions virtually independent of one another. Rather, energy conver-
Researchers worldwide are recognizing the promise of μgrids sion, heat recovery and use, and renewable harvesting may all
to provide heterogeneous SQRA to serve sensitive loads, to be taking place simultaneously at various locations within the
improve energy efficiency by moving thermal generation close building energy system. Second, the structure of energy flows
to possible uses. This would permit waste heat recovery and in the building must accommodate multiple energy processes
use, and better integration of small-scale dispersed renewables in a manner that permits high overall efficiency. In other words,
into the energy supply infrastructure. Nonetheless, it is also the building must be designed around its energy flows and en-
clear that development of μgrid concepts and capabilities will ergy equipment to ensure efficiency. And third, multiple quali-
require considerable RD3 resources, and efforts are currently ties of electricity may be supplied to various building functions,
underway, in Europe, North America, and Japan, intended to and there placement and supply must be considered.
demonstrate μgrid concepts, operation, and economic viability. DER-CAM, developed as part of the CERTS Microgrid RD3
Close cooperation and exchange of information among these programme is intended to permit economic analysis of possi-
disparate activities can deliver the most efficient RD3 agenda ble building μgrids. DER-CAM finds the optimal combination
overall. The international Microgrids Symposiums held so far of equipment to install in a building-scale μgrid, given the re-
have offered a highly beneficial forum for exchange of relevant quirements for useful energy services in the building, the local

ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT! 39

1,074 MARNAY, FIRESTONE PANEL 1. THE FOUNDATIONS OF A FUTURE ENERGY POLICY

economic environment, e.g. utility tariffs, and a menu of avail- Acknowledgements

able equipment. This approach allows a quite new type of build- The work described in this report was funded by the Office of
ing energy analysis that directly delivers a desirable equipment Electricity Delivery and Energy Reliability, Distribution System
choice taking the potential interactions between heat (includ- Integration Program of the U.S. Department of Energy under
ing cooling) and on-site generating equipment into account. Contract No. DE-AC02-05CH11231, and by the California En-
ergy Commission, Public Interest Energy Research Program,
References under Work for Others Contract No. 500-03-024. The author
recognizes the significant contributions of the group that in
Amin, Massoud, 2005, Powering the 21st Century: We Can
part organized the Microgrids Symposiums and have provided
and Must Modernize the Grid, IEEE Power & Energy
much of the material used in this paper, Hiroshi Asano, Univer-
Magazine, March-April 2005.
sity of Tokyo, Nikos Hatziargyriou, National Technical Univer-
Amin, Massoud and Bruce Wollenberg, 2005, Toward a Smart
sity of Athens, and Reza Iravani, University of Toronto. The au-
Grid, IEEE Power & Energy Magazine, September-Octo-
thor also thanks the many participants in the three completed
ber 2005.
Micrigrids Symposiums. The paper also builds on prior work
Engler, Alfred, 2006/2007, Applicability of Droops in Low
the author has completed together with Giri Venkataramanan,
Voltage Grids, International Journal of Distributed
University of Wisconsin, Madison, Afzal S Siddiqui, University
Energy Resources, special issue on microgrids, Decem-
College, London, Michael Stadler, Center for Energy and Inno-
ber-January.
vative Technologies, Yspertal, Austria, plus Kristina Hamachi
European Commission, 2006, European Smart Grids
LaCommare, and Judy Lai of Berkeley Lab.
Technology Platform, Vision and Strategy for Europe’s
Electricity Networks of the Future, April.
Funabashi, Toshihisa, and Ryuuichi Yokoyama, 2006, Micro- Glossary
grid Field Test Experiences in Japan, paper presented at AC alternating current
the IEEE Power Engineering Society General Meeting, Berkeley Lab Ernest Orlando Lawrence Berkeley National
Montréal, June. Laboratory, Berkeley CA
Gellings, Clark, Marek Smotyj, and Bill Howe, 2004, The CERTS Consortium for Electric Reliability Technology
Future’s Smart Delivery System: Meeting the Demands for Solutions
High Security, Quality, Reliability, and Availability, IEEE CHP combined heat and power
Power & Energy Magazine, September-October. CM CERTS Microgrid
Hatziargyriou, Nikos (ed.), 2006/2007, International Journal DC direct current
of Distributed Energy Resources, special issue on microg- DER distributed energy resources
rids, December-January. DER-CAM DER Customer Adoption Model
Hirose, Keiichi. T. Takeda, and S. Muroyama, 2006, Study on genset traditional reciprocating engine powered
Field Demonstration of Multiple Power Quality Levels generator
System in Sendai, International Telecommunications ISET e.V. Institut für Solare Energieversorgungstechnik,
Energy Conference, Rome, September. Kassel, Germany
Katiraei, Farid, and Reza Iravani, 2007, Power Management kV kilovolt
Strategies for a Microgrid with Multiple Distributed Gen- kW kilowatt
eration Units, paper TPWRD-00620-2005, accepted for MCFC molten carbonate fuel cell
publication in the IEEE Transactions on Power Systems. MW megawatt
Lasseter, Robert, Abbas Akhil, Chris Marnay, John Stevens, NAS sodium-sulphur (battery)
Jeff Dagle, Ross Guttromson, A. Sakis Meliopoulos, Rob- NEDO New Energy and Industrial Technology
ert Yinger and Joe Eto, 2003, Integration of Distributed Development Organisation, RD3 branch of the
Energy Resources: The CERTS Microgrid Concept, CEC, Japan’s economy ministry
P500-03-089F, October. NREL National Renewable Energy Laboratory,
Lasseter, Robert, 2006, Dynamic Distribution Using (DER): Golden CO
Distributed Energy Resources, paper presented at the NTT Nippon Telegraph and Telephone Corp., the
IEEE Power Engineering Society T&D Meeting, Dallas, dominant Japanese telecom
May. NTUA National Technical University of Athens
Lynch, Jonathan, V. John, S.M. Danial, E. Benedict, I. Vi- O&M operating and maintenance
hinen, B. Kroposki, and C. Pink, 2006, Flexible DER Util- PAFC phosphoric-acid fuel cell
ity Interface System: Final Report, National Renewable PCC point of common coupling
Energy Laboratory, NREL/TP-560-39876, August. PV photovoltaic
Marnay, Chris and Giri Venkataramanan, 2006, Microgrids RD3 research, development, demonstration, and
in the Evolving Electricity Generation and Delivery Infra- deployment
structure,” paper presented at the IEEE Power Engineer- SMA SMA Technologie AG, a power electronics
ing Society General Meeting, Montréal, June. manufacturer based in Niestetal, Germany
SOFC solid oxide fuel cell
SQRA security, quality, reliability, and availability

40 ECEEE 2007 SUMMER STUDY • SAVING ENERGY – JUST DO IT!