policymaking for microgrids: economic ... - building … · the economics of microgrids arises from...

66 IEEE power & energy magazine may/june 20081540-7977/08/$25.00©2008 IEEE

©E

YE

WIR

E

Digital Object Identifier 10.1109/MPE.2008.918715

TTECHNICALLY, MICROGRIDS ARE EMERGING AS AN OUTGROWTH OF DISPERSEDon-site and embedded generation via the application of emerging technologies, especially power elec-tronic interfaces and modern controls, and, similarly, microgrid economic and regulatory analysis isgenerally rooted in the same approaches used to evaluate distributed energy resources (DER). As in theeconomics of many traditional on-site generation projects, the economics of heat recovery and its appli-cation by combined heat and power (CHP) systems is central to the evaluation of microgrids, and inte-gration of this capability is a key requirement whenever CHP appears as an option. The recovery ofwaste heat offers a key advantage to generation close to loads but at the same time adds significantly toanalysis complexity because of the need to simultaneously meet requirements for electricity and heat,plus the inevitability of storage, both active and passive, entering the equation. More novel is the eco-nomics of power quality and reliability (PQR), which in microgrids can potentially be tailored to therequirements of end uses in a manner only considered to a limited degree in utility-scale system; e.g.,by interruptible tariff options. The economics of microgrids arises from evaluation methods for on-sitegeneration from the customer perspective and from the traditional utility economics of expansion plan-ning from the utility perspective. Both of these areas have received considerable attention, so a growingtoolkit exists, but methods need reinforcement in some key regards. Central to public policymaking willbe consideration of the societal impact of microgrids, especially since their adoption may changemacrogrid requirements. While partially explored, this topic is still in need of rigorous analysis.

Setting the Scene

Big Picture EconomicsCost-reflective pricing mechanisms are critically important to facilitating competition in generation. Inthe context of the impact on network costs, this is manifested in the cumulative value chain from

may/june 2008 IEEE power & energy magazine 67

power generation to consumption. Electricity produced by largecentral generation is sold in European wholesale markets foraround 3–5 EURcent/kWh, but by the time this electricity reach-es the end consumer it is being sold at a retail price of 10–15EURcent/kWh. This increase in value is driven by the added costof transmission and distribution services to transport electricityfrom the point of production to consumption. Note that most ofthese costs occur at the point of interface between the macrogridand the customer through metering and billing, customer servic-es, storm recovery, etc. Despite its critical effect on economics,this point is often overlooked in discussions of the relative effi-ciency and cost of small- versus large-scale generation, the prac-tical limitations on the possible beneficial application of CHP,and the high cost and low power density of renewable resources.Conversely, it is often argued that if utility customers are free toself-generate or not at will, they are not paying the full costs oftheir burden on the macrogrid, as is further discussed below.

One of the central conceptual promises of microgrids is thatmultiple decision-making responsibilities may be concentrated inthe hands of one agent, thereby removing one of the seriouscauses of market failure in existing centralized power systems,most clearly evident in underinvestment in energy efficientequipment. In other words, imagine a single decision makerbeing responsible for purchasing commercial energy inputs, gridpower, generator fuels, and generating equipment, and at thesame time being responsible for choosing end-use equipment,storage technology, and potentially taking advantage of any localopportunity fuels such as solar or a bio waste stream. Perhapsthis omnipotent decision maker will trade off these complexalternatives in the evenhanded manner that has escaped usheretofore, resulting in chronic underinvestment in energy effi-ciency. For example, with a payback of months in some applica-tions, compact fluorescent light bulbs still command only about5% of the U.S. market. Similarly, small-scale renewables andstorage technologies have been slow to penetrate the market inpart because they often lie beneath the radar screens of utilityplanners. Successfully combating climate change will require usto overcome these past failings, and microgrids offer one promis-ing opportunity for achieving a wiser future. Part of the analysischallenge, therefore, is to develop methods and tools that can

guide developers of microgrids toward desirable decision-making at both the design and operatingstages. Further, the desirable outcomes might involve complex combinations of technologies andbehaviors. While economic methods for evaluation of on-site electricity supply and CHP are welldeveloped, areas in need of research attention remain. In general, past analysis of CHP has relied onrather simplistic technical rules of thumb and inadequate consideration of the importance of complexmarket incentives such as time-of-use or feed-in tariffs. More specifically, microgrids bring to thefore the issues of waste-heat-driven cooling, on-site energy storage, and heterogeneous PQR, whichare all relatively uncharted areas of engineering-economic analysis.

Importance of Building CoolingThe growth of building cooling requirements with its conse-quent effect on peak load growth, system investment require-ments, and electricity supply reliability has been an area ofconcern in most parts of North America and Japan for sometime; increasingly the warmer areas of Europe face similarconcerns. In Greece, the summer peaks in 2005, 2006, and2007 exceeded their corresponding winter peaks by approxi-mately 10%, 15%, and 20%, respectively, and in 2004 and2006, Spain experienced winter peaks only about 5% largerthan its summer peaks. Further, if climate change brings occa-sional hot summers, such as France experienced in 2003,spurts of air conditioning adoption may severely tax macro-grids. Further, France’s heat wave simultaneously caused lowriver flows that reduced hydro generation and limited nuclearoutput because of reduced cooling capacity. These develop-ments enhance the argument for local provision of electricityservice to vital services and the use of CHP systems to pro-vide building cooling, which lowers peak electrical loads. Thispossibility poses some new challenges because it creates asimultaneous effect in which the recovery of waste heat fromon-site generation lowers the requirement for generation and,in fact, for capacity along the whole supply chain. Further,cooling of buildings can be achieved using a host of approach-es: passive cooling, traditional electrically powered compres-sor air conditioning, direct-drive equivalents, absorption cycletechnology using either direct fire or waste heat, passive meth-ods, precooling using building thermal mass, etc. Designingmicrogrids such that waste heat is effectively utilized poses achallenge in general, and the addition of this cooling compli-cation significantly adds to the burden. Note, however, theimportance of waste heat recovery from the energy efficiency(or, equivalently, climate change) perspective. Currently, thewaste heat emitted from U.S. power plants is responsible forapproximately 28% of the energy-related carbon emissions ofthe country. This is of the same order as all transportation sec-tor emissions, which currently represent about 32%. Further,the waste-heat-related emissions are growing as electricitybecomes a more dominant fuel, and more of the generationmix is thermal generation, which means that if current trendscontinue, transportation will be overtaken by waste heat as asource of carbon emissions in five years or so.

StorageEconomic evaluation of electrical, thermal, or fuel storagealways poses a complex problem because of its inter-temporal

nature, i.e., the way storage is operated in any time step affectsits operation in many others, and because of the need to makedecisions despite uncertainty surrounding future circum-stances. Deployment of electrical storage on large scales islimited, e.g., in pumped hydro stations, although combustionfuels are stored in massive quantities, e.g., in natural gas tanks.The waste heat from thermal power generation is almost neverstored on a large scale. Microgrids bring all these aspects ofthe storage problem together. If load variation, supply inter-mittency, or hard economics require it, local electrical storageis a promising option for microgrids, particularly given theemergence of new battery technologies, such as flow batteries.Because, in most instances, heat and electrical loads do notoccur together, heat storage may also be attractive. For exam-ple, an office building may need considerable electricitythroughout the day but space heating primarily during themorning hours when occupants arrive for work and/or coolingduring hot afternoon hours. Storage may be active using spe-cific devices, such as fluid tanks, or may be passive, such asby precooling or heating of a building to take advantage of itsthermal lag. Additionally, because a key objective of micro-grids may frequently be to provide reliable power during gridoutages and because these may accompany disruptions in fuelsupplies, on-site storage of fuel(s) may also be desirable.

Power Quality and ReliabilityAn important aspect of microgrid economics that has minimalevaluation methods available is local control of PQR. The tra-dition in electricity supply has been one of universal standardsfor PQR. While not necessarily achieved in practice, the para-digm for control of PQR has been to strive for the same stan-dard in all parts of the supply network at all times. Achievingthe targets has incurred both physical and operational costs;i.e., maintaining and improving PQR require both investmentin equipment and redundancy but also conservative operatingprocedures that preclude some economic transactions, withconsequences costly to societies. Microgrids create an oppor-tunity for a radical rethinking of this paradigm. By sophisticat-ed local control by a microgrid, the prospect of tailoring PQRto match the requirements of end-uses becomes a promisingsource of economic gain. Note that many end-use loads needonly low PQR, and the ones that do require gourmet powerare typically small. Matching PQR requirements more pre-cisely than is done today might yield considerable benefits.Further, multiple PQR supplies might be delivered to variousparts of a single device. Analysis approaches to determining

68 IEEE power & energy magazine may/june 2008

The economics of microgrids arises from evaluation methods for on-site generation for the customer perspective and from traditional expansion planning for the utility perspective.

may/june 2008 IEEE power & energy magazine

the desirable level of PQR either globally in the macrogrid orlocally in the microgrid simply do not exist.

The Complex Regulatory EnvironmentThe regulatory environment into which microgrids are enter-ing is extremely complex, in part because they encroach onmultiple areas of existing regulation not conceived with themin mind, i.e., generator interconnection rules, air quality per-mitting, building codes, tariffication, etc.

Traditional interconnection rules required generators totrip in the event of any disturbance, while one of the keyobjectives of microgrid development is to achieve systemsthat can island and ride through grid problems, or in someproposals partially so. Clearly, a conflict exists. Through con-siderable effort in IEEE committees, proposed rules areemerging to overcome the problem. Operating parts of thedistribution network or customer sites in island mode, i.e.,isolated from the upstream network, with the voltage and fre-quency regulated locally by resources outside the jurisdictionof the utility may also create conflicts with existing codes andregulations. Safety and damage liability issues also may arisesince the distribution utility may no longer be responsible forthe operation of the island.

In general, market mechanisms are still not mature enoughto accommodate the envisaged participation of microgridentities. Where feed-in tariffs exist, they are usually fixed forsmall-scale DER, often based on the primary energy source,while the retail consumer kWh tariffs in many countries arecentrally regulated. Metering issues also need to be resolved,since current regulatory practice requires metering energy atthe connection point of each individual user, rather than at theaggregation point of user groups.

A European ViewFundamental to development of European energy policy hasbeen the move toward liberalized competitive markets thatsupport and promote the development of cost-effective futuresystems, together with a significant emphasis on climatechange policy that requires cutting greenhouse gas emissionsfrom the power sector. There is a clear focus in the EuropeanUnion on promoting low-carbon generation technologies andrenewables with a new and binding target of a 20% renew-able portfolio standard by 2020. Over the last decade, thispolicy commitment has been matched with national incentiveschemes and support mechanisms for renewable and low-car-bon distributed energy sources that have contributed to areduction in technology costs and an increasing penetrationof smaller-scale generation into distribution networks.

In this context, more recently there has been an increasedinterest in various forms of microgeneration, primarily insmall-scale and domestic CHP and photovoltaics (PV). InGermany, more than 1.5 GW of PV has been installed in res-idential and commercial sectors as a result of favorable tar-iffs. Another interesting example is that the Woking BoroughCouncil installed the first local-authority-private-wire CHP

system in the United Kingdom, providing heating to civicoffices, a local parking lot, two hotels, and leisure centers.Woking installed 1.5-MW and 950-kW CHP systems withthermal storage and heat-activated absorption coolingtogether with a 200-kW fuel cell with CHP and a number ofPV panels. The system can island, and is operated by anenergy service company that is involved in the operations,the investment in primary energy plant (boilers and chillers)and their eventual replacement, electricity and heat distribu-tion, and supply of green energy direct to customers. Suchdevelopments have in turn raised the question of (in)appro-priateness of the existing technical, commercial, and regula-tory framework to support cost-effective integration of these

emerging generation technologies into overall system opera-tion and development. Based on Woking’s success, the Lon-don Climate Change Agency has recently been established toaccelerate reductions in London’s greenhouse gas emissions.The new agency aims to deliver low- and zero-carbon energyprojects and services involving a combination of energy effi-ciency; combined cooling, heating, and power; renewables;and other innovative technology including waste-, water-,and transport-related projects. One of the showcase projectsis London City Hall (Figure 1), which features 617 PV mod-ules with peak capacity of 52.4 kW installed on the domedroof and 46 bespoke translucent glass-glass laminates withpeak capacity of 14.6 kW forming an “eyelash” that pro-vides solar shading for the ninth floor (London’s LivingRoom). There are over 28,000 individual PV cells with atotal active area of 417 m2. The system is expected to gener-ate about 50,000 kWh of electricity each year, i.e., 1.5% ofCity Hall’s current total energy consumption, and save over28,000 kg of CO2 each year.

Economics

Microgrid Perspective: A U.S. ExampleAs mentioned above, the economics of microgrids from thecustomer perspective arises out of the evaluation of on-siteand embedded generation. However, the participants in a

69

figure 1. London City Hall.

microgrid are not customers in the traditional sense; rather,they are a grouping of coordinated sources and sinks that areoperating in unison, so they appear to the utility as a singleentity, either a net source or net sink. Coordination here goesbeyond simply acting as a group. The microgrid is closer toan integrated energy system designed and operated to meetthe energy requirement (electrical, heating, and cooling) of itsmembers as well as to supply electricity of PQR commensu-rate with the requirements of end-use loads.

One notable research effort in the United States to devel-op methods for the evaluation of microgrids is the Distrib-uted Energy Resources Customer Adoption Model(DER-CAM), which has emerged from the Consortium forElectric Reliability Solutions (CERTS) research programfunded by the U.S. Department of Energy (USDOE) and theCalifornia Energy Commission (CEC). It identifies optimal

technology-neutral DER investmentsand operating schedules at a givensite, based on available DER equip-ment options and their associated cap-ital and O&M costs, customer loadprofiles, energy tariff structures, andfuel prices. The Sankey diagram inFigure 2 shows partially disaggregatedmicrogrid end uses on the right-handside and energy inputs on the left.DER-CAM solves this entire problemoptimally and systemically, findingthe combination of equipment andtheir operating schedules that meet theload requirements. The goal functionis typically cost minimization, but car-bon minimization or other objectivesare possible.

Figure 3 and Figure 4 show someexample results from DER-CAM for ahypothetical San Francisco, California,hotel with 25,000 m2 of floor space.

This is a contrived example in which storage has been madeavailable at low cost to demonstrate the potential complexityof solutions. The optimal system chosen for this site consistsof a 200-kW natural-gas-fired reciprocating engine (gasengine) with heat recovery, a 722-kW solar thermal array thataugments engine heat recovery, a 585-kW absorption chiller,1,100 kWh of electrical storage, and 299 kWh of heat stor-age. In Figure 3, the pink line shows the true hourly heatdemand of the hotel, which is only for water heating on thissummer day. The black line shows the heat load including theheat used for cooling, which is the difference between thetwo curves. When total heat production goes above the blackline, storage is being charged, while the purple colored arearepresents heat recovery from storage. Note that heat is sup-plied from two sources: CHP heat recovery from the gasengine and, during the daytime, solar thermal collection. Fig-ure 4 shows the electricity balance on the same day. Similar-ly, the pink line shows actual electricity requirements, whilethe black line is the actual requirement plus electricitydemand that has been displaced by absorption cooling as rep-resented by the dark blue striped area. Note that a consider-able fraction of electricity is still purchased from the utility,although the generator runs at close to its maximum of 200kW for most of the day. The electricity storage is used to sup-plement supply during the expensive daytime hours (1200 to1800 are the local on-peak tariff hours).

Microgrid Perspective: A Japan ExampleA similar approach to DER-CAM’s has been applied tosome potential microgrid host sites in Japan. Three micro-grid hosts are assumed, as shown in Table 1; the first one(No. 1) includes a mixed office and apartment building; thesecond one (No. 2) is a hospital, also with apartments; and


Sun

Utility Electricity TotalElectricity

Electricity Only End Uses

Refrigeration andBuilding Cooling

Hot Water

Building Heating

Natural Gas Only End Uses

Natural Gas

WasteHeat

Ther

mal

Stor

ageRecovered

heat

On-SiteGeneration

UtilityNatural

Gas

figure 2. Energy flows in a microgrid from fuels to end uses.

0

200

400

600

800

1,000

1,200

Hour

The

rmal

Pow

er (

kW)

Heat Load Heat and Abs. Chiller Load

CHP Heat

ThermalStorage

Solar Thermal

1 5 9 13 17 21

Storage Charging

figure 3. Heat balance for a July weekday (source:CERTS).


the third (No. 3) consists of a hotelwith office and retail space. The peaksummer demand of all three hosts isabout 1 MW.

This work has found that the optimalnumber of gas engines selected is deter-mined by the ratio between baseload andpeak electricity demand together withthe ratio between heat and electricitydemand. As the ratio of baseload to peakincreases, the number of gas enginesselected decreases because larger unitshave higher energy conversion efficiency.Technological advances in gas engineshave been significant—overall thermalefficiency is now at least 40% (lowerheating value), i.e., higher than the typi-cal 350-kW class gas engine currentlyused in power generation systems. Fig-ure 5 shows the scale economy of energyconversion efficiency (y-axis) of a gasengine (plotted at load factors of 50%,75%, and 100%).

Some battery storage is required tosupply electricity for a limited time dur-ing the peak demand period, and a ther-mal storage tank for space cooling andspace heating is selected to minimize theuse of a direct natural-gas-fired absorp-tion chiller. Figure 6 shows the resultsfor a winter day.

The same research has considered thebeneficial effects of customer load aggre-gation. For example, when a hotel, retailstore, and office operate their own gasengines, it requires five units, and anaverage of 32.5% is achieved. If loads areaggregated, the total number of units canbe reduced to two, and average generat-ing efficiency is improved to 37.6%,while annual cost is reduced by 4.5%.Conversely, depending on a smaller num-ber of units reduces redundancy andtherefore reliability, which creates a com-plex trade-off in actual applications.

Utility Perspective:An Unlikely BeginningInterestingly, at least one outstandingexample of the initial innovative concep-tual thinking about a more dispersedelectricity supply system came from theutility perspective. Around 1990,researchers at the Pacific Gas and Elec-tric (PG&E) research laboratory in San

71

0

200

400

600

800

1,000

1,200

Hour

Ele

ctric

ity (k

W)

Noncooling Electric Load Total Electric Load

Purchase

Generate

Absorption Cooling Offset Storage

Storage Charging

Storage Charging

1 5 9 13 17 21

figure 4. Electricity balance for a July weekday (source: CERTS).

Ratio of Off-PeakDemand to On- Ratio Between

Total Floor Peak Demand: Heat and ElectricityNo. Configuration Area [m2] Electricity Demand1 Office 25,000 0.350 1.01

Apartment 42,0002 Hospital 29,000 0.278 1.57

Apartment 20,5003 Hotel 8,000

Retail 3,000 0.213 1.18Office 20,000

table 1. Host configurations for case studies.

42.0

39.0

36.0

33.0

30.0

27.0

24.00 200 400 600

Engine Scale [kW]

800 1,000 1,200

Load Factor (L.F.) 100%

L.F. 75%

L.F. 50%

Estimated Eff. (Lower Than 350 kW) (L.F. 100%)Estimated Eff. (Higher Than 350 kW) (L.F. 100%)

Gen

erat

ing

Effi

cien

cy [%

]

figure 5. Scale economy of generating efficiency of a gas engine (source: University of Tokyo).

Ramon, California, began to explore the possibility of usingPV deployment as a way to postpone expansion of the distribu-tion system, especially substations. The investigators observedthat the load on the distribution infrastructure is highly peakywhere air conditioning is prevalent, and all elements of the sup-ply chain must be sized to meet the peak. In fact, in California,air conditioning contributes about 14% to the state’s annualpeak, while it is only responsible for about 5% of the electricityconsumption. Further, in the hot interior parts of the state,peaks occur on hot sunny afternoons when PV output could beconfidently relied upon. Using fortuitously available data for a

500-kW PV installation planned for an area near a saturated10.5-MW substation, the benefit-cost ratio of such an installa-tion to delay necessary system upgrades was estimated.

At the time, the PG&E work found that the PV option wasnot competitive, but its researchers’ innovative thinking kickedoff a new line of reasoning that has ultimately led to question-ing of the fundamental logic of the entire centralized powersystem, most notably in developing economies where the lega-cy grid is weak or nonexistent. Locally in northern California,methods for valuing dispersed deployment of renewables insome local distribution networks have now reached a highstage of development. One study has developed a method forassessing the benefits of distributed generation (DG) sys-temwide for some local utilities. The relevance of this body ofwork is in part that the value of sources or net load reductionsin the current supply system could provide a revenue streamfor microgrids if they are independent entities and incentives toutilities to develop microgrids themselves. Current pricing ofelectricity, as discussed below, typically tends to be constantover large areas, but this consistency is an artifact of the regula-tory process and does not deliver accurate price signals to cus-tomers in most cases. Particularly, the added costs of delivering

power to congested areas and those where the infrastructure isreaching limits are not reflected in prices.

Utility Perspective: Avoided Cost EstimationMicrogeneration, located close to demand, delivers electricitydirectly with limited requirement for use of the network. Thispower may therefore have a higher value than that of conven-tional generation due to the potential of DG to reduce thedemand for distribution and transmission network capacity,reduce losses, and, if integrated, increase the reliability of sup-ply seen by the end customers. The importance of understand-

ing and quantifying these benefits hasbeen recognized, although these are yetto be incorporated within the technical,commercial, and regulatory framework.Clearly, ignoring these particular fea-tures in the derivation of the value ofmicrogeneration results in noncompeti-tive markets in which small-scale gen-eration cannot compete on a levelplaying field with conventional genera-tion. Ultimately, this will createinefficient and non-cost-reflective sys-tems, whereby microgeneration is notefficiently integrated into the system,and the resulting framework relies onunnecessary network reinforcementand inefficient solutions based onincreasingly expensive, low-utilizationconventional generation.

The potential benefits and costs ofintegration of PV and micro CHP intosystem operation and development

will be driven by a number of factors including:! level of penetration! density (distribution)! correlation between generation operation patterns

and demand profiles.Regarding the impact on networks, given that microgener-

ation is located near load, this will result in reduction of loss-es and release of network capacity, which can be used todefer future network reinforcement needed to accommodateload growth.

In Southern Europe for example, where peak demandoccurs on hot summer days, PV generation is likely to reducelosses in distribution networks. In this context, it is importantto remember that losses are a quadratic function of load, and somost losses occur in summer daytime as this is the most heavi-ly loaded time for the majority of networks. This coincideswith the expected operation of PV, and the loading on the net-work and losses will therefore be reduced. The results of stud-ies carried out on some typical networks for different levels ofpenetration are presented in Figure 7.

The reduction of losses is expressed as a percentage of thebase losses that are the distribution losses with no PV


2,500

2,000

1,500

1,000

500

0

kWh/

h

0.666666667 0.833333333 1 1.666666667 1.333333333 1.5Time of Day

Heat Release from Thermal StorageSteam Recovered from Gas Engines

Gas Absorption Refrigerator

Demand

figure 6. Space-heating demand and supply for a winter weekday (source: University of Tokyo).


installed in the system. It can be concluded that PV can makea very significant contribution to loss reduction; especially inlong rural distribution networks (up to 40%).

In Northern Europe, on the other hand, there has beenincreased interest in the network benefits of the application ofmicro CHP. Again, the operation of CHP tends to coincide withpeak demand that occurs on winter evenings. Studies carriedout on typical U.K. distribution networks show similar trends:the reduction in network losses driven by micro CHP couldreach between 25% in urban and 40% in rural systems, which isclearly very significant. On the other hand, installation of PV inNorthern Europe will tend to have less beneficial impact givenwinter-driven peak demand. The estimated maximum lossreduction could be between 10% and 20%.

Regarding the benefits from realizing network capacity byreducing peak demand, studies on the U.K. distribution net-work suggest that the value of savings in future network rein-forcement and replacements enabled by domestic CHP couldbe worth up to about !150 /kW of microgeneration connected.Furthermore, increased reliability of electricity supply createdby microgrids that are capable of operating in an islandingmode could be significant given that faults in medium-voltagedistribution networks are responsible for the vast majority ofinterruptions seen by end consumers. Studies conducted in theUnited Kingdom indicate that the economic benefit of the reli-ability improvement would be about !30 per domestic con-sumer (and significantly more for commercial customers).

Another important benefit of domestic CHP is in its contri-bution to reduction of the amount of conventional generationcapacity required to meet system peak demand, given that theoutput is typically coincident with thewinter peak. This effectively reducesthe demand at the winter peak seen bythe rest of generation, thus reducingthe capacity required to maintain sys-tem security. Averaged across manyCHP units, this effect would be fairlypredictable and reliable, so domesticCHP could actually displace existingconventional generation capacity.Given that, typically, energy displacedby domestic CHP is smaller than theconventional capacity displaced, thiscreates savings that are estimated at 2!c/kWh produced by CHP.

However, turning these economicbenefits created by microgrids into

commercial income will require significant changes in the regu-latory framework, and this transition is in its infancy.

Societal PerspectiveConsiderable analytic efforts have been made to portray thefull societal balance sheet for distributed power systems, butconsiderable challenges remain nonetheless. An outstandingexample is the 2002 book Small Is Profitable by AmoryLovins. This book represents a heroic effort to collect all theinformation pertinent to development of a dispersed systeminto one volume and make the definitive case for a distrib-uted system.

Another notable effort to describe a few of the societalbenefits of a dispersed power system was produced by theUSDOE last summer. Among discussion of other benefits, thestudy reports on outstanding examples of DG providingimproved resilience to vital services. For example, the Missis-sippi Baptist Medical Center was able to operate as an islandfor 52 hours following Hurricane Katrina using its on-site 3.2-MW turbine. It was the only hospital in the area able to main-tain full-service operation. Under normal operatingconditions, the system provides 80% of total electricity, 95%of steam, and 75% of cooling using two absorption chillers.

RegulationThe issue of interconnection pervades all sizes of independentgeneration, but it becomes most problematic for relativelysmall-scale operations, for which the cost of complex engineer-ing studies and/or special equipment would be prohibitive.Here DER more generally have blazed the trail for microgrids;

73

If load variation, supply intermittency, or hard economics require it,local electrical storage is a promising option for microgrids,particularly given the emergence of new battery technologies.

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

100%91%79%68%57%45%34%23%11%0%

PV Installed Capacity in % of Peak Load

Loss

Red

uctio

n (%

)

UrbanRural

figure 7. Reduction in losses as a function of PV generation penetration.

however, microgrids raise some particular technical issues.Most notably, they are intended to operate as islands and mayadditionally be controlling PQR semi-autonomously, whereasmany interconnection rules require DER to disconnect in theevent of a disturbance.

Interconnection: EuropeIn Europe, DER interconnection rules are typically set at thenational level by macrogrid technical regulations, regulatordirectives, or national standards. Regardless of the normativeprocedures applied, interconnection practices aim to ensurethat DER will not disturb other users of the network in nor-mal operation, and that safety will not be jeopardized in caseof abnormal conditions. To this end, interconnection proce-dures typically include technical provisions on the following:! voltage regulation and power quality, including steady-

state voltage deviations, fast variations, flicker, harmon-ics, dc injection

! power factor! protection and anti-islanding schemes! earthing-grounding arrangements.

Although country-specific variations do exist, intercon-nection practices and rules invariably treat DER as a potentialsource of disturbance for the network, an attitude related toand originating from the traditional unidirectional operatingparadigm of radial distribution networks. This attitude is bestreflected in the anti-islanding provisions of all existing Euro-pean codes, which force immediate disconnection duringblackout to prevent potential safety threats to other networkusers and utility field crews, as well as to avoid operation andprotection complexities. Anti-islanding is realized via passiveor active protection schemes. Passive anti-islanding protec-tion typically utilizes voltage and frequency relays at theinstalled DER terminals that determine whether island condi-tions exist. Active anti-islanding capabilities common ininverter-connected DER are based on more sophisticated androbust algorithms for detecting loss of grid conditions. In allcases, DER are forced to quickly disconnect (in times rang-ing from a tenth of a second to a few seconds), precluding thepossibility of DER supporting its local part of the grid. Oncea DER unit has disconnected from the network, it cannotreconnect into a deenergized system.

Although current grid codes do not permit the formationof islands within the distribution network, they do permitintentional islanding of user facilities; i.e., the separation of aspecific private installation from the public network to

improve its own PQR. This can be accomplished if the userinstallation includes suitable DER resources and the intercon-nection is designed to provide backup islands during networkoutages. This approach requires, besides sufficient and reli-able embedded generation to support local load, careful coor-dination with utility sectionalizing and protection equipment.Upon occurrence of upstream network faults or outages, theinstallation isolates from the network, forming an island,which is supported by local generation (and possibly stor-age), self-providing frequency, and voltage regulation. Toreturn from island to grid-connected mode, suitable synchro-nizing equipment is required. Such an approach is compatiblewith microgrids, which typically apply intentional islandingand controlled reconnection of one self-controlled microgridor of one comprising several coordinated users.

To establish a framework suitable for the development andproliferation of microgrids in Europe, the overall intercon-nection framework needs to be revised, taking the followingspecific issues into account.

! During islanding, micro sources will operate in a con-fined network with drastically altered characteristicswhile performing active regulation and control duties.

! Deliberate islanding of network sections without lossof local generation should be permitted while penaliz-ing unintentional (and potentially harmful) islanding.Note that the overcurrent protection typically installedat the connection point of every network user may nolonger be effective in the islanded mode of operationdue to the low fault currents expected.

! The standardization of network components and con-struction will also need to change for sections of the net-work where microgrid operation is foreseen. This wouldrequire the installation of controllable interconnectionbreakers and paralleling means at the point of couplingof the whole microgrid, e.g., at the low-voltage busbarsof a distribution substation, where fuses would normallysuffice. Modified network protection devices might berequired within a microgrid section, along with specificprovisions for neutral earthing upon isolation from themain grid. Suitable communication infrastructure run-ning along with the power lines might also be needed. Ingeneral, distribution utilities will have to modify theiroverall network development practices where the forma-tion of microgrids is foreseen.

! Finally, metering arrangements will also need to bemodified. The installation of a simple meter per user

IEEE power & energy magazine may/june 200874

Turning these economic benefits created by microgrids into commercial income will require significant changes in theregulatory framework, and this transition is in its infancy.


installation will no longer suffice. Additional meteringof incoming and outgoing power and energy will berequired at the overall microgrid point of common cou-pling, while more sophisticated metering tools willdefinitely be required to support the participation of themicrogrid in markets.

Interconnection: JapanIn Japan, the Ministry of Economy, Trade, and Industry(METI) has established rules and technical guidelines for gridinterconnection. As amended in 2004, these rules regulatesafety and assure power quality in Japan’s highly reliable net-work. METI’s rules establish technical standards for electricequipment under the Electricity Industry Law. Requirementsinclude relays, switches for protection, islanding prevention,and communication systems. The technical guidelines forpower quality govern interconnection in the low-voltage(100–200 V), high-voltage (6–22 kV), and extra-high-voltage(> 22 kV) networks and include voltage regulation limits;e.g., 101 ± 6 V, 202 ± 20 V for low-voltage distribution linesand voltage dips under 10% are limited to 2 s. However, theserules contain no provisions on cost bearing among parties,which may cause problems. For example, excess voltage out-side the required range is expected after the first connectionbetween a distributed generator and the distribution line goessmoothly. The grid connection might be based on a first-come, first-served basis and cannot produce social-optimumpenetration of distributed generators. Furthermore, it is neces-sary to unify the standards required for interconnecting dis-tributed generators with information and distributionnetworks. Implementing control and protection of distributedgenerators as good citizens that needs a two-way communi-cation system between distributed generators and the systemoperator.

Interconnection: The United StatesEarly U.S. research showed that interconnection barriers are sig-nificant and costly. Released in 2000, one influential study inter-viewed developers who had attempted to interconnect newtechnologies at customer sites. Of the 65 projects reviewed, onlyseven were interconnected in a timely manner. Concern aboutbarriers to distributed power has led to considerable efforts toenable lower hassle and cost of interconnection. As with manyU.S. regulatory issues, interconnection is seriously complicatedbecause utility regulation resides, in large part, at the state level.Within the United States, therefore, the progress of interconnec-tion regulation is quite uneven. Certain states have moved ahead,which in some cases has involved legislative action.

In 2003, after five years of development, the IEEE 1547Standard for Interconnecting Distributed Resources withElectric Power Systems was published with the goal of creat-ing a set of technical requirements that could be used by allparties on a national basis. During IEEE 1547 development,it was recognized that islanding parts of the distribution sys-tem, depending on implementation, could improve reliability

or, if unplanned, be a serious problem. The concepts of inten-tional and unintentional islanding were developed anddefined in the standard. IEEE 1547 addresses unintentionalislanding by requiring the DR to detect the island conditionand cease to energize the area electric power system (EPS)within 2 s of island formation. Although it was realized thatintentional islanding (microgrids) could be beneficial to dis-tribution system and customer reliability, this issue wasdeferred for consideration in future revisions of the standard.In late 2003, IEEE formed a new project under the IEEE1547 series to address intentional islanding and microgridscalled IEEE P1547.4 Guide for Design, Operation, and Inte-gration of Distributed Resource Island Systems with ElectricPower Systems. Currently in draft, this document will covermicrogrids and intentional islands that contain DR connectedat a facility level and with the local utility. It provides alterna-tive approaches and good practices for the design, operation,and integration of the microgrids and covers the ability toseparate from and reconnect to part of the utility while pro-viding power to the islanded local power systems. The guidewill cover the DR, interconnection systems, and participatingelectric power systems. It is intended for use by designers,operators, system integrators, and equipment manufacturers.Its implementation will expand the benefits of DR byenabling improved EPS reliability and building on therequirements of IEEE 1547.

75

figure 8. Hachinohe street showing parallel public and private distribution feeders.

Macrogrid-Utility Interaction at a Demonstration in JapanAn indicative problem between the microgrid and its localmacrogrid has been experienced during a New Energy and Indus-trial Technology Development Organization (NEDO)-fundedmicrogrid demonstration project in Hachinohe, northern Hokkai-do. This project began operation in October 2005 and is beingevaluated for PQR, cost effectiveness, and carbon-emissionsreduction over its demonstration period, which ended in March2008. The microgrid has PV and wind turbines totaling 100 kW,510 kW of controllable gas engines supplied by digester gas froma sewage plant, and a 100-kW lead-acid battery bank. Seven

Hachinohe City buildings are supplied via a private 6-kV, 5.4-kmdistribution feeder (shown in Figure 8), with the whole systemconnected to the commercial grid at two nearby points of com-mon coupling at City Hall and its annex building. Test islandingoperation was also demonstrated at this project.

Figure 9 shows a schematic of the project, and Figure 10shows the main site. This water treatment plant contains thebiogas digestor with waste-wood-fired heat, the biogasengines, and a battery bank.

Under an agreement with the local utility company, electrici-ty can serve the City Hall and annex buildings, the largestmicrogrid users, but is not allowed to flow into the private distri-

bution line at internal (not common) couplingpoint C, so a reverse power relay is installedthere. The under-power relay at couplingpoint A between the commercial grid andHachinohe City Hall and the reverse powerrelay at coupling point C have a reaction timeof 1 or 2 s. Unfortunately, the relays may acti-vate in the event of fluctuations caused by thestarting or stopping of large electrical equip-ment during low demand periods, especiallyat night. This situation has created a costlyhard constraint on operation of the microgrid.As in many microgrid demonstrations inJapan, a constant power flow at the couplingpoint was a major objective of the Hachinohedemonstration and has been successfullyachieved, although by application of complexmicrogrid controls. In this case, a harsh


figure 10. Central generation facility at the Hachinohe water treatment plant.

Hachinohe City Hall Hachinohe City HallAnnex

CB1 CB2 CB3 CB4 CB4

CB5 CB7Load (Electricity)PV2 WTs

The Project SystemThe Commercial Grid

InternalCoupling Point C

Private Distribution Line

Load (Electricity)

CommonCoupling Point A

CommonCoupling Point B

TOBU DrainageFacility(Sewage Plant)Load (Heat)3 Gas EnginesBatteryPV5 Boilers

OtherDemandsLoad(Electricity)2 PVs2 WTs

figure 9. Main loads in the Hachinohe microgrid and its macrogrid interconnection.


requirement was placed on the microgrid with serious economicconsequences, so further deregulation of electricity marketsmight reduce the risk of these operational constraints. Analysisof the control strategy will be carried out to determine its impacton operating cost. A more economic control strategy would cer-tainly have reduced the necessary battery and reserve capacities.

Environmental RegulationEven though small-scale power generation accounts for a tinyshare of urban pollution compared to mobile sources, since thetwo are sometimes regulated differently, microgrids might beseverely disadvantaged. This is the case in California whereonly about 3% of urban nitrogen oxide and minimal hydrocar-bon emissions are derived from any form of generation, withthe overwhelming majority of emissions coming from vehicles.Yet, small generators are soon to be subject to very strict stan-dards, set statewide for microturbines and fuel cells but locallyfor gas engines, and further, California’s legislated objective ofreducing the state’s greenhouse gas emissions to 1990 levels by2020 stands in conflict with prevailing urban air quality rules,which are likely to impede the development of microgrids thatcould improve the utilization of fossil fuels overall.

The division of authority in California between station-ary sources and mobile sources has caused a major imbal-ance in the weight of regulation on the two pollutionsources. One possible form of redress would be to regulatebased on overall footprint rather than stack emission rates.A microgrid may involve thermal generation, but at thesame time, use of CHP, renewable generation, and efficientend-use devices may mean that the total footprint of themicrogrid is lower than the macrogrid alternative. If stan-dards were set on this basis, a more desirable overall out-come might be achieved.

ConclusionsSome aspects of microgrids represent a radical departurefrom existing approaches to provision of electricity, whileothers are outgrowths of existing systems, or even returns toways of doing business commonplace to the industry duringbygone times. Developing economic and regulatory meth-ods to accommodate them will require us to look in manyplaces for inspiration and tools. In some areas a solid basisis in place while elsewhere we still face challenges.

AcknowledgmentsMuch of the work described in this article has been sponsoredby NEDO, the European Commission, USDOE, and the CEC.

For Further ReadingR.B. Alderfer, T.J. Starrs, and M.M. Eldridge, “Makingconnections: Case studies of interconnection barriers andtheir impacts on distributed power projects,” NationalRenewable Energy Laboratory, NREL/SR-200-28053,May 2000.

Connection Criteria at the Distribution Network for Dis-tributed Generation, CIGRE TF C6.04.01, Feb. 2007.

Energy and Environmental Economics, Inc., and Elec-trotek Concepts, Inc., “Renewable distributed generationassessment: Alameda Power and Telecon Case Study,”prepared for the California Energy Commission PublicInterest Energy Research program, no. 500-20050010,Jan. 2005.

Y. Fujioka, H. Maejima, S. Nakamura, Y. Kojima, M.Okudera, and S.Uesaka, “Regional power grid with renew-able energy resources: A demonstrative project in Hachino-he,” in Proc. CIGRE 41, Paris, 2006, no. C6-305.

A.B. Lovins, Small is Profitable: The Hidden EconomicBenefits of Making Electrical Resources the Right Size.,Snowmass, CO: Rocky Mountain Institute, 2002.

D. Sugar, R. Orens, A. Jones, M. El-Gassier, and A.Suchard, “Benefits of distributed generation in PG&E’stransmission and distribution system: A case study of pho-tovoltaics serving Kerman substation,” PG&E R&D Report007.5-92.9, Aug. 1992.

U.S. Department of Energy (2007, June), “The potentialbenefits of distributed generation and rate-related issues thatmay impede its expansion,” [Online]. Available:http://www.oe. energy.gov/epa_sec1817.htm

U.S. Department of Energy and U.S. EnvironmentalProtection Agency (2007, Nov.), “Aligning utility incen-tives with investment in energy efficiency,” [Online].Available: http://www.epa.gov/cleanrgy/documents/incen-tives.pdf

BiographiesChris Marnay is a staff scientist at Berkeley Lab and leads itsEnd-Use Forecasting Group.

Hiroshi Asano is currently a professor at the University ofTokyo and a senior research scientist with the CentralResearch Institute of Electric Power Industry.

Stavros Papathanassiou is an assistant professor at theNational Technical University of Athens.

Goran Strbac is a professor at Imperial College Londonand a director of the U.K. Centre for Distributed Generationand Sustainable Electrical Energy.

77

As with many U.S. regulatory issues, interconnection is seriously complicated because utility regulation resides, in large part, at the state level.

p&e

policymaking for microgrids: economic ... - building … · the economics of microgrids arises from...

Documents