Integrating distributed generation into electric power ...distributedwind.org/assets/2013/05/Integrating-distributed... · Integrating distributed generation into electric power ...

A

ooopanapio©

K

1

(sotat

t(riWi

0d

Electric Power Systems Research 77 (2007) 1189–1203

Integrating distributed generation into electric power systems:A review of drivers, challenges and opportunities

J.A. Pecas Lopes a,∗, N. Hatziargyriou b, J. Mutale c, P. Djapic c, N. Jenkins c

a INESC Porto, Portugalb NTUA, Greece

c The University of Manchester, UK

Available online 9 October 2006

bstract

It is now more than a decade since distributed generation (DG) began to excite major interest amongst electric power system planners andperators, energy policy makers and regulators as well as developers. This paper presents an overview of the key issues concerning the integrationf distributed generation into electric power systems that are of most interest today. The main drivers behind the focus on DG integration, especiallyf the renewable type, in many countries around the world are discussed. A synopsis of the main challenges that must be overcome in the process isresented. Particular emphasis is placed on the need to move away from the fit and forget approach of connecting DG to electric power systems topolicy of integrating DG into power system planning and operation through active management of distribution networks and application of otherovel concepts. The paper also analyses the repercussions in transmission system operation and expansion that result from the connection of largemounts of DG of different energy conversion systems focusing on issues related with impacts in steady state operation, contingency analysis,
rotection coordination as well as dynamic behaviour analysis. A discussion on the possibility of provision of ancillary services by DG is alsoncluded. Some results from studies performed in the interconnected Portuguese transmission system are presented and discussed. Some of thepportunities that could be exploited in support of the integration and hence greater penetration of DG into electric power systems are also explored.
2006 Elsevier B.V. All rights reserved.

at and

tptw

gaoaap(S

eywords: Distributed generation; Distributed energy resources; Combined he

. Introduction

Although it is now over a decade since distributed generationDG) became a significant and topical phenomenon in powerystems, there is as yet no universal agreement on the definitionf DG, which is also known as embedded generation or dis-ributed generation. Current definitions of DG are very diversend range from 1 kW PV installations, 1 MW engine generatorso 1000 MW offshore wind farms.

Nowadays it is more common for DG to be considered inhe context of the wider concept of distributed energy resourcesDER), which includes not only DG but also energy storage andesponsive loads. The power system architecture of the future,
ncorporating DER, will look very different from that of today.
hilst the pace of change is likely to be evolutionary, the changetself is expected to be nothing short of a revolution as many

∗ Corresponding author.E-mail address: [email protected] (J.A.P. Lopes).

s2leo2a

378-7796/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.epsr.2006.08.016

power

raditionally held views and approaches to system operation andlanning developed over the last 100 years are challenged andransformed to suit the requirements envisaged in the brave neworld of the future.This revolution has already commenced as evidenced by the

rowth of DG worldwide ([1]; [2]), as Governments strive tochieve ambitious targets of incorporating considerable amountsf distributed renewable generation (DRG) and combined heatnd power (CHP) in response to the climate change challengend the need to enhance fuel diversity. For example, the Euro-ean Union in the White Book on Reneable Energy SourcesRES), adopted in 1998 ([3]—White Paper for a Communitytrategy and Action Plan), has set as target 12% of electricityupplied by renewable generation by 2010. According to the000 Directive of the European Parliament [4], this is trans-ated to an electricity production of 22.1% from renewable
nergy sources. In particular, the progress in wind power devel-pment in recent years is impressive. In autumn 2002, almost7.260 MW of electricity-generating Wind Turbines are oper-ting in fifty countries. Of these, about 75% (20.280 MW) are
mailto:[email protected]

dx.doi.org/10.1016/j.epsr.2006.08.016

1 System

imttaehtdtc

pttDiaoaa

softdfa

2

rtnT

2

2

sieqtiia

2c

slo

rnsomsw

aaIg

2

2s

cbtipwps

2q

fiona

2

2s

ereipE

ai

-

190 J.A.P. Lopes et al. / Electric Power

nstalled in the European Union, with leading countries coveringore than 89%, Germany, Spain and Denmark [5,6]. In addi-

ion, considerable developments have been recently made onhe technological front, and potentially most challenging in thebove respect, is the development of micro-turbines and novelnergy storage technologies. The increasing penetration of DGas several technical implications and opens important ques-ions, as to whether the traditional approaches to operation andevelopment of power systems are still adequate. This is par-icularly true at the distribution level, where the bulk of DG isonnected.

This paper discusses the main drivers behind DG growth andresents an overview of the challenges that must be overcome inhe integration of DG into electric power systems. In particular,he need to move from the fit and forget policy of connectingG to electric power systems to a policy of integrating DG

nto power system planning and operation through active man-gement of distribution networks is emphasised. Some of thepportunities that could be exploited to support the integrationnd hence greater penetration of DG into electric power systemsre also discussed.

The paper also analyses the repercussions in transmissionystem operation and expansion that result from the connectionf large amounts of DG of different energy conversion systemsocusing on issues related with impacts in steady state opera-ion, contingency analysis, protection coordination as well asynamic behaviour analysis. Some results from studies per-ormed in the interconnected Portuguese transmission systemre presented and discussed.

. Drivers of DG growth

The primary drivers behind the growth of DG and the cur-ent focus on its integration into electric power system opera-ion and planning can be classified into three main categories,amely environmental, commercial and national/regulatory.hese drivers are discussed briefly below.

.1. Environmental drivers

.1.1. Limiting green house gas (GHG) emissionsThe use of renewable energy and CHP to limit GHG emis-

ions is one of the main drivers for DG. In this regard, it ismportant to point out that integration of renewable sources oflectrical energy into power systems is a somewhat differentuestion from that of the integration of DG into power sys-ems. Integration of DG includes some of the issues related tontegration of renewable sources but clearly does not deal withntegration of transmission connected renewable sources suchs large on shore and off shore wind farms.

.1.2. Avoidance of the construction of new transmissionircuits and large generating plants

Another important driver for DG from the environmental per-pective is the avoidance of construction of new transmissionines and large power plants to which there is increasing publicpposition. There is however also opposition from some envi-

-

s Research 77 (2007) 1189–1203

onmental lobby groups to on-shore wind farms on grounds ofoise and visual “pollution”. There is therefore a balance to betruck between the need for sustainable energy solutions on thene hand and the need to maintain scenic beauty of the environ-ent. Some argue that environmentally benign technologies,

uch as wind, that do not emit any GHG and have no long termaste management problems should be favoured.Technological developments in generator technology are

lready delivering cost effective small to medium size gener-tion technologies for domestic application such as micro-CHP.n times when a premium is placed on land use these technolo-ies are likely to prove popular.

.2. Commercial drivers

.2.1. General uncertainty in electricity markets favoursmall generation schemes

One of the acknowledged consequences of the introduction ofompetition and choice in electricity is the increased risk facedy all players in the electricity supply chain from generatorshrough transmission and distribution businesses to retailers. Its well known that the capital outlay required to establish newower stations can be very high. The uncertainties associatedith a competitive market environment may favour generationrojects with a small capacity whose financial risk is commen-urately small.

.2.2. DG is a cost effective route to improved poweruality and reliability

The presence of DG close to load centres can have a bene-cial impact on power quality and supply reliability. One areaf improvement is voltage profile improvements, reduction inumber of customer minutes lost especially if DG is allowednd able to stay on when there are network outages (islanding).

.3. National/regulatory drivers

.3.1. Diversification of energy sources to enhance energyecurity

In recent times, there has been increasing concern amongstnergy policy makers regarding energy security. There is aecognition that modern societies have become so dependent onnergy resources to the extent that should there be a disruptionn its supply the consequences would be too ghastly to contem-late in political, economic and social terms. Because of this theU energy policy focuses on energy security and sustainability.

In the context of energy security and sustainability, DG isn attractive proposition in many respects. Some of the moremportant ones are for example:

It is distributed around the network close to customers—failure of one power station will have limited impact on thewhole system compared to failure of one large power plant or
bulk electricity transmission facility.Diverse technologies and primary energy sources—by diver-sifying the energy sources especially utilising renewablesources there is sense of control over the nation’s future energy

System

2

egattsaap

3

ir

3

3

an

3

idlqngTtilsls

3

tfctptnn

3

ssweHsbts

3

botnfimoatcad

-

-

-

J.A.P. Lopes et al. / Electric Power

needs. There is increasing concern that the bulk of fossil fuelbased energy supplies come from regions of the world wherecontrol of these resources could be potentially unpredictablethus posing an unacceptable risk.

.3.2. Support for competition policyThere is an abiding faith amongst the proponents of reform of

lectricity supply industries that introduction of competition ineneration and customer choice will deliver low energy pricesnd better service quality. One of the prerequisites for effec-ive competition to occur is that there must be many players inhe market. DG clearly advances this cause by providing manymall generators that could potentially trade in the energy marketnd, where appropriate market arrangements exist, also trade inncillary services. The opportunities for DG to participate in therovision of ancillary services are discussed later in this paper.

. Challenges to increased penetration of DG

Challenges to increased penetration of DG can be classifiednto three main categories, namely technical, commercial andegulatory. These are discussed in turn below.

.1. Technical

.1.1. Voltage rise effectThe voltage rise effect is a key factor that limits the amount of

dditional DG capacity that can be connected to rural distributionetworks.

.1.2. Power qualityTwo aspects of power quality are usually considered to be

mportant: (1) transient voltage variations and (2) harmonicistortion of the network voltage. Depending on the particu-ar circumstance, DG plant can either decrease or increase theuality of the voltage received by other users of the distributionetwork. Power quality is an increasingly important issue andeneration is generally subject to the same regulations as loads.his tends to work well in practice and it is generally possible

o meet the required standards by careful design. The effect ofncreasing the network fault level by adding generation ofteneads to improved power quality. A notable exception is that aingle large DG, e.g. a wind turbine, on a weak network mayead to power quality problems particularly during starting andtopping.

.1.3. ProtectionA number of different aspects of DG protection can be iden-

ified: Protection of the generation equipment from internalaults; protection of the faulted distribution network from faulturrents supplied by the DG; anti-islanding or loss-of-mains pro-ection (islanded operation of DG will be possible in future as
enetration of DG increases) and impact of DG on existing dis-ribution system protection. All these aspects are important andeed to be carefully addressed in connecting DG to distributionetworks.
wwn

s Research 77 (2007) 1189–1203 1191

.1.4. StabilityTraditionally, distribution network design did not need to con-

ider issues of stability as the network was passive and remainedtable under most circumstances provided the transmission net-ork was itself stable. Even at present stability is hardly consid-

red when assessing renewable distributed generation schemes.owever, this is likely to change as the penetration of these

chemes increases and their contribution to network securityecomes greater. The areas that need to be considered includeransient (first swing stability) as well as long term dynamictability and voltage collapse.

.2. Commercial

Case studies have indicated that active management of distri-ution networks can enable significant increases in the amountf DG that can be connected to the existing networks. Althoughhe cost associated with the operation of active distributionetworks is still to be identified, it is expected that the bene-ts are likely to considerably outweigh the cost of its imple-entation. At present, however, distribution companies that

perate wires businesses have no incentives to connect DGnd offer active management services. In order to supporthe development of active distribution networks and extractorresponding benefits associated with connecting increasedmount of DG, new commercial arrangements need to beeveloped.

Generally, three approaches are possible:

To recover the cost of implementing active managementdirectly through the price control mechanism (increasing theamount of recoverable capital and operating expenditure asso-ciated with active management). The cost recovery could beachieved through increased charges for the use of the net-works (imposed either to distributed generators benefitingfrom active management and/or demand customers).To establish an incentive scheme that would reward compa-nies for connecting DG, such as one recently developed inthe UK [7]. Such an incentive mechanism, assuming a suit-able design of the scheme, could lead to the development ofactive distribution networks. Such schemes could be fundedfrom increased charges imposed on generators and/or demandcustomers.To establish a market mechanism, outside of the regulatoryframework, which would create a commercial environmentfor the development of active networks. Under this scenario,distribution companies would offer active management ser-vices to generators for a charge. Clearly, whenever the netbenefit from active management exists, this could be used asa basis for bilateral negotiations between the local companyand the generator.

The development of active management of distribution net-orks could stimulate further unbundling of distribution net-ork services with an exchange of services between distributionetwork operators and distributed generators.

1 System

3

itphtao

4m

tpc

vmwsotpt

dsTcdElptfl

Fn

ff(ehripu(hpr

daprtca

dcomoemigacm


.3. Regulatory

In the absence of a clear policy and associated regulatorynstruments on the treatment of DG, it is very unlikely that thisype of generation is going to thrive. The reasons for this areartly historical and related to the way distribution networksave been developed and operated as passive networks. In ordero foster the required changes, there is a clear need to developnd articulate appropriate policies that support the integrationf DG into distribution networks.

. Path forward—active distribution networkanagement

Active distribution network management is seen as the keyo cost effective integration of DG into distribution networklanning and operation. This is in direct contrast to the currentonnect and forget approach.

The historic function of “passive” distribution networks isiewed primarily as the delivery of bulk power form the trans-ission network to consumers at lower voltages. These networksere designed through deterministic (load flow) studies con-

idering the critical cases so that distribution networks couldperate with a minimum amount of control. In other words con-rol problems were solved at the planning stage. This practise ofassive operation can limit the capacity of distributed generationhat can be connected to an existing system.

For well-designed distribution circuits, there is little scope foristributed generation when simple deterministic rules (e.g. con-ideration of minimum load and maximum generation) are used.his practice significantly limits the connection of DG. As theseonditions may only apply for a few hours per year it is clearlyesirable to consider stochastic voltage limits, as proposed underuropean standard EN 50160. The application of probabilistic

oad flow [8,9] and Monte-Carlo simulation techniques providerobabilities of voltage limit violations and thus leads to objec-ive decision-making. In Fig. 1, example of a probabilistic loadow analysis is provided [10] showing the probability density

ig. 1. Impact of wind farm output on the frequency distribution of voltages atode.

sdba

Fmtals

s Research 77 (2007) 1189–1203

unctions of the voltage at the connection bus of a Wind Parkor low levels of installed capacity. For low levels of penetration0.675 MW), the voltage was generally below 1.0 per unit. How-ver, at higher levels of penetration (8.0 MW), the likelihood ofigh voltages increased. With a high penetration, the voltageise effect is very clear (15% above nominal). It is, however,mportant to note that whilst the voltage rise is pronounced, therobability of its occurrence is low, less than 6% in this partic-lar case. Using crude studies, connecting a larger wind farm8 MW) may have been rejected without considering the likeli-ood of the increase in voltage. No attention would have beenaid to the fact that the mean voltage level over time would beelatively unaffected by the size of the wind farm.

This analysis would also allow distributed generation toecide to be constrained off in certain circumstance to limit volt-ge rise. Further, many DGs have the ability to operate at variousower factors and may even be able to act as sources/sinks ofeactive power when not generating. For some overhead distribu-ion circuits (i.e. those with high reactance) then the DGs couldontribute to circuit voltage control provided suitable controlnd commercial systems were in place.

In contrast, active management (AM) techniques enable theistribution network operator to maximise the use of the existingircuits by taking full advantage of generator dispatch, controlf transformer taps, voltage regulators, reactive power manage-ent and system reconfiguration in an integrated manner. AM

f distribution networks can contribute to the balancing of gen-ration with load and ancillary services. In future, distributionanagement systems could provide real-time network monitor-

ng and control at key network nodes by communicating withenerator controls, loads and controllable network devices, suchs reactive compensators, voltage regulators and on load taphanging-transformers (OLTC). State estimation and real-timeodelling of power capability, load flow, voltage, fault levels and

ecurity could be used to make the right scheduling/constrainingecisions across the network. These techniques will probablye applied gradually rather than fulfilling all the above listedttributes right from the beginning (Bopp et al., 2003).

The DMS controller software has two functional blocks (seeig. 2): state estimation and control scheduling. The state esti-ation block uses the network electrical parameters, network

opology, load models and real-time measurements to calculatenetwork state estimate. The measurement input comprises the

ocal and network measurements. This is passed to the control-cheduling block, which uses it to calculate a new set of control

Fig. 2. Block diagram of DMS controller software.

System

vcot

evs

---

co

----

vra

eoes

nDs

5

issdat

D

mtdeat

a

fieCatotbii

a

------

5

tffrebOr(

fnspt

bwtuvr

vD

5

a


alues for the devices connected to the network. The set ofontrol values optimises the power flow in the network, whilstbserving all the constraints and taking account of all the con-racts.

The network data input comprises the network topology andlectrical parameters. The pseudo measurement input providesalues for all un-measured quantities that are required for thetate estimator.

The constraints input comprises three types of constraint:

Primary plant constraints, e.g. distributed generator capacity.Control limits, e.g. OLTC maximum number of tap steps.Network constraints, e.g. voltage limits.

The contracts input comprises the details of ancillary serviceontracts between the distribution network operator and ownersf:

distributed generators;controllable loads;reactive power compensators;automatic voltage control relays.

Coordinated voltage control with on load tap changers andoltage regulators generally leads to operating arrangementsequiring a number of measurements from key network pointss well as communications.

From a regulator’s perspective, active management shouldnable open access to distribution networks. It has the functionf facilitating competition and the growth of small-scale gen-ration. In addition, the use of the existing distribution assetshould be maximised to minimise costs to consumers.

Therefore, an integral understanding of the interrelated tech-ical, economic and regulatory issues of active management andG is important for the development of the future distribution

ystems [11].

. Ancillary services from DG

As DG penetration increases it will become an economicmperative that DG participates in the provision of ancillaryervices needed for secure and reliable operation of the powerystem. This is important for the simple reason that if DG onlyisplaces the energy produced by central generation but not thessociated flexibility and capacity, the overall cost of operatinghe entire system will rise.

Another reason for exploring ancillary service provision byG is to improve the economic viability of some DG projects.There is potential for distribution network ancillary service

arkets to develop in-line with the anticipated increase in elec-ricity generation from distributed resources. A study was con-ucted to investigate this potential [12]. The study sought tovaluate the distribution ancillary service market opportunities
pplicable to both renewable and non-renewable forms of dis-ributed generation.
A pre-requisite for the detailed development of operationalnd commercial models was that any new services should be

cf

s

s Research 77 (2007) 1189–1203 1193

nancially material to the distributed generator whilst remainingconomically and operationally attractive to network operators.onsequently, value based approaches were adopted for eachncillary service with a view towards improving the attrac-iveness of distributed and renewable generation projects. Inrder to demonstrate the application of the proposed approachhe UK electricity market was used as a case study partlyecause it is one of the most advanced competitive marketsn Europe and provides practical examples of the issues undernvestigation.

The services for which potential arrangements were exploredre:

TSO frequency response;TSO regulating and standing reserve;TSO reactive power;DNO security of supply contributions;DNO quality of supply services;DNO voltage and power flow management services.

.1. TSO frequency response

Frequency response services are required by the TSO to main-ain the system frequency within statutory tolerances. A keyeature of TSO frequency response provision is the requirementor generators to be part loaded. It is unlikely that TSO frequencyesponse services will be provided regularly by renewable gen-ration, as the rewards from providing such services woulde insufficient to compensate for losses of ROC (Renewablebligation Certificate, a green energy premium) and energy

evenues. Distribution connected combined cycle gas turbineCCGT) plant already provides this service to TSOs.

Infrastructure and generator size considerations suggest thatrequency response services will remain attractive to large,on-renewable generators. An innovative, low-cost means ofcheduling automated mass responses from highly flexible smalllant can be developed. If such a scenario does arise, it is likelyhat the scope for aggregation services will increase.

Although mandatory frequency response capabilities mayecome a technical requirement for large distribution connectedind farms, thereby ‘resolving’ any infrastructure constraints,

he extent to which the TSO will utilise such capabilities isnclear. The attractiveness to wind operators, of regularly pro-iding such services is similarly unclear due to the loss of ROCevenue.

The value of TSO frequency response is estimated toary between D 0.59/kW per annum for wind generation and3.72/kW per annum for CCGT technology.

.2. TSO regulating and standing reserve

Reserve energy is required to provide rapid access to gener-tion, to accommodate errors in demand forecasting, to provide
ontingency arrangements for generation failures and to restorerequency response capabilities.
The key differences between frequency response and reserveervices relate to delivery timescales. Typically, reserve services

1 System

aqtr

rrtaop

5

estt

iTtliDmwt

ccobuHs

5

ctacmiv

bmTb

vtd

Ie

act“fc

5

sotatt

trcs

5

prtfronD

ndsbwat

5t

Taot


re manually initiated and involve longer lead times. A conse-uence of simplified service initiation procedures is a reduc-ion in the sophistication of control requirements, thus makingeserve more attractive to smaller providers.

It is unlikely that synchronised reserve will be provided byenewable generation, as the compensation for the loss of ROCevenue (green energy premium) would be insufficient to jus-ify part load operation. Non-renewable distributed generationlready provides standing reserve services to the TSO at a valuef approximately D 10.4/kW per annum. Increased DG partici-ation could be facilitated by expanded aggregation services.

.3. TSO reactive power

TSO reactive power can be sourced from distributed gen-rators, especially those connected at 132 kV, for transmissionystem voltage regulation. Reactive power sourced at lower dis-ribution voltages will effectively displace reactive power fromransmission-connected generation.

DG connected at lower voltage levels can make a significantmpact on the amount of reactive power exchanged betweenSO and DNO systems. A simple generic model was developed

o illustrate DNO reactive power import reductions at differentevels of DG penetration. The value of DG derived reductionsn DNO reactive imports was estimated to be approximately1.78/kW per annum. The impact of DG on TSO reactive powerarket will be driven by many different variables. More workould be required to determine the impact of DG on DNO reac-

ive requirements.The impact of reactive power management on the transport

apabilities of distribution circuits was also investigated. DGonnected close to loads could extend the transport capabilitiesf existing circuits. The value of this service would be limitedy the low cost of power factor compensation equipment. It isnlikely that this would represent significant income for DG.igh DG availabilities would be needed for DNOs to consider

uch services.

.4. DNO security of supply contributions

The proposed planning recommendation ER P2/6 in the UKould broaden opportunities for DNOs to consider contributionso network security from DG. However, as most DNO networksre largely ER P2/5 compliant, the requirements for securityontributions from DG may be limited in the short-term. In theedium to long term, load growth and asset replacements could

ncrease opportunities for DG to provide network support ser-ices.

The value of security provided by non-intermittent DG cane related to the avoided or deferred costs of network reinforce-ent. DG can also substitute for network automation facilities.his is particularly relevant when considering security contri-ution of intermittent generation such as wind.

A number of examples were used to illustrate the potentialalue of network security services. For non-intermittent genera-ors, values in the range of D 1.49–17.83 kW−1 per annum wereerived, depending upon the complexity of the network solution.

esd

s Research 77 (2007) 1189–1203

t is anticipated that most reinforcements would be at the lowernd of this range.

Because of the drive to reduce customer interruptions (CI)nd customer minutes lost (CML), DNOs in the UK have madeonsiderable investments in 11 and 0.4 kV networks. A result ofhis investment is that distribution networks in UK are generallyover compliant” with planning and security standards. For theoreseeable future, the scope for DG to provide security servicesould be limited.

.5. DNO quality of supply services

In the future, there could be opportunities for DG to improveervice quality on 11 and 0.4 kV networks, given the contributionf such networks to quality of supply statistics. In order for DGo improve service quality on such networks, the generation mustlso be connected at 11 or 0.4 kV, thus restricting opportunitieso relatively small sized generation. A key requirement for DG,o reduce the impact of outages, is islanded operating capability.

Analysis suggests that the annual benefit of islanding opera-ion was approximately D 2.08 and D 28.23 kW−1 per annum foresidential and commercial customers, respectively. Due to theomplexity of islanding, it is unlikely that DG will be able toignificantly reduce CIs and CMLs in the short or medium term.

.6. DNO voltage and power flow management services

Analysis revealed that voltage control and flow managementroblems are essentially network planning related issues as theyelate to supply restoration times following network failures (inhe context of UK security standards ER P2/5 or P2/6). Theocus here is in generators providing services critical to systemestoration (e.g. voltage support or flow control) after a faultccurs rather than under normal operating conditions when theetwork will tend to provide services to DG in order to maximiseG output (e.g. tap changing and flow control).Because of the relatively low availability of DG compared to

etwork components and the UK’s deterministic voltage stan-ards, opportunities will be limited for DG to provide voltageupport or overload reduction. Only non-intermittent DG woulde suitable for such applications. Opportunities will improveith increased DG penetrations due to the higher collective

vailability. The value of these services was estimated to be ofhe order of D 2.23/kW per annum.

.7. Ancillary service capabilities of different generationechnologies

Whilst all of the above services were explored in detail, onlySO frequency response, TSO regulating and standing reservend DNO security of supply contributions represent realisticpportunities for distributed generators in the short or mediumerm.

Combined cycle gas turbines (CCGT) and DFIG wind gen-rators were the most promising technologies for the provi-ion of TSO frequency response services whereas CCGTs,iesel standby generators and perhaps micro-CHP were best

System

paadec

st

ttirt

aIEte

stmt

ui

bsDci

lnr

itntpdsw

flsv

il

t

ci

aidaepirsipt

fs

6o

ucw(

ichtbeTpw

otc

--

6

cgfl


laced to provide reserve services. Synchronous wind gener-tors connected to the grid through power electronics interfacelso present good capabilities for frequency control. Moreoverue to pitch control, some squirrel-cage induction wind gen-rator may also contribute (in a limited way) to frequencyontrol.

It was found that, to varying degrees, DNO security of supplyervices could be provided by most existing distributed genera-ion technologies.

As the majority of existing DG has been installed for elec-ricity supply purposes, very few generators are equipped withhe infrastructure necessary to provide ancillary services. Suchnfrastructure includes governors, automatic voltage regulators,esynchronisation facilities, and appropriate protection, moni-oring and communication facilities.

The most appropriate commercial arrangements for responsend reserve services appear to be market-based mechanisms.deally the TSO’s current arrangements could be extended.xpanded aggregation arrangements, utilising lower cost infras-

ructure, would facilitate increased participation from small gen-rators.

The most appropriate commercial arrangements for DNOecurity of supply services appear to be bilateral contracts dueo the local and site-specific nature of security requirements. A

ethodology for pricing distribution network use that recogniseshe contribution of DG to security has been illustrated.

Opportunities for DG to provide ancillary services willndoubtedly increase as DG penetrations and availabilitiesncrease.

The analysis undertaken suggests that value of the most feasi-le ancillary services will be relatively low. Consequently, suchervices will represent incremental revenue opportunities forG. In general, it would not be possible to develop business

ases for investing in DG solely on the basis of ancillary servicencome.

Niche opportunities will emerge for DG to provide ancil-ary services, usually in circumstances where constraints restrictetwork development, e.g. environmental, planning and terrainelated constraints.

In an era with significantly increased levels of DG operat-ng on active distribution networks, the opportunities for DGo provide ancillary services may increase. However, on activeetworks there is an increased likelihood that due to distribu-ion network constraints, certain modes of operation may not beermitted by the DNO. Consequently, there could be increasedelivery uncertainty regarding the provision of TSO ancillaryervices from distributed generators connected to active net-orks.In circumstances where a distributed generator receives con-

icting instructions regarding the provision of different ancillaryervices, local services should take precedence over national ser-ices.

Higher penetrations of DG will increase DNO options regard-
ng network operation and development decisions, which couldead to lower overall costs.
Increased penetration of DG could also enhance competi-ion in TSO markets for frequency response and reserve. This

aubb

s Research 77 (2007) 1189–1203 1195

ould be particularly relevant should demand for these servicesncrease with intermittent generation.

In the UK energy market under new electricity tradingrrangements (NETA), the output from distributed generations largely purchased by suppliers and settled through supplieremand accounts. Therefore, suppliers must ensure they areware of generator operating regimes and also whether gen-rator operation is likely to be influenced by ancillary servicerovision. Supplier concerns will relate to imbalance exposuresn the BM and the fulfilment of ROC targets. Suppliers willequire notification of ancillary service provision, in order touitably revise demand forecasts. Ancillary service instructionsssued post-Gate Closure will inevitably impact upon a sup-lier’s imbalance exposure and potentially reduce the value ofhe energy supplied.

It is important to stress that provision of ancillary servicesrom DG will not jeopardise or degrade security of supply buthould contribute to its enhancement.

. Technical impacts of DG on transmission systemperation

The technical impacts of DG in system operation can be eval-ated at different levels: (a) steady state operation; (b) protectionoordination; (c) dynamic behaviour (with a natural interactionith the protection solutions adopted, namely from the DG side);

d) provision of ancillary services.Evaluations of the impacts in the transmission network

nvolve a large number of studies, considering different typi-al scenarios of operation (for instance peak load hours, valleyours and mid consumption levels). In power systems wherehere is a considerable contribution from hydropower, the num-er of these scenarios increases since it will be necessary tovaluate impacts for winter (wet or dry) and summer periods.he intermittent nature of some of DG plants (such as windarks) can also lead to the need to consider additional scenariosith different levels of DG production.These impacts can be analysed by comparing the solutions

btained for each scenario before and after the introduction ofhe DG levels for a given study. Two types of analyses can beonducted:

Global evaluation of system performance.Analysis of the electrical behaviour in specific network areas,due to specificities of the energy conversion systems adoptedand due to the characteristics of the local network.

.1. Steady state operation

Steady state operation impacts are evaluated in terms ofhanges in voltage profiles, active and reactive losses and con-estion levels in system branches through the solution of loadow problems. A change in the amount of DG level also requires
previous power dispatch in order to allocate to the conventionalnits the power that needs to be produced for each scenario toe considered. Assuming that the total DG production level cane defined as PDG and the total consumption level is PL, the

1196 J.A.P. Lopes et al. / Electric Power Systems Research 77 (2007) 1189–1203

ortug

ai

P

o

s

Fig. 3. Single line diagram of the P

mount of power to be produced by the conventional units (PG)s:

G = PL + Plosses − PDG

This means that the steady state impacts are largely dependentn the final dispatch solution to be adopted.

dc2s

uese transmission system in 2000.

An example of the type of results obtained from this type ofteady state impacts studies is presented next, through the results
erived from the analysis performed in the Portuguese inter-onnected transmission grid, described in Fig. 3 (Pecas Lopes,002). Three transmission levels were taken into account in thistudy: 150, 220 and 400 kV. The three interconnections corridors

J.A.P. Lopes et al. / Electric Power Systems Research 77 (2007) 1189–1203 1197

ive lo

wu

yoio

atLslvdtoigtiiedttito

ampnais

tmlhipe

tro

ofo

attb

cedd

dtiw

-

-

tft

eflptar

Fig. 4. Changes in transmission act

ith Spain were also considered in this analysis, by keeping thesual import/export plan on each of the analysed scenarios.

The impact evaluation study was conducted in order to anal-se the influence of an additional connection of about 300 MWf DG in the north part of Portugal, without reinforcing the exist-ng network. An amount of 1300 MW, with large participationf cogeneration, was considered to be already in operation.

The Portuguese system is characterized by a considerablemount of hydro production in the north and a large concentra-ion of thermal production in the area of Lisbon and south ofisbon. Six scenarios were therefore considered for this analy-is: peak load wet winter (PHI), peak load dry winter (PSI), peakoad dry summer (PSV), valley hours during wet winter (VHI),alley hours during dry winter (VSI) and valley hours duringry summer (VSV). For the solution of the load flow problem athe transmission level, DG plants can be modelled as injectionsf active and reactive power. This modelling approach, adoptedn the described study, results also from the fact that, in Portu-al, these plants are obliged to follow a generation rule such thathey have to produce, during the off-valley hours, reactive powern the amount of 40% of the active power, and they should notnject reactive power in the grid during the valley hours. How-ver, in the case of large concentration of induction generators aifferent bus bar model needs to be considered in order to tacklehe impacts of voltage on active and reactive power outputs ofhe machines. The R, X model (or special P, Q model) describedn Ref. [13] can be used to provide more accurate results fromhese types of studies, and to account for near to voltage collapseperating conditions.

In network areas with large concentration of wind gener-tion using asynchronous machines, large voltage excursionsay happen also as a result of the amplification effect in reactive

ower produced by shunt capacitor banks used to generate theeeded reactive power, which varies with the square of the volt-ge [14]. In these areas if an excessive low voltage takes place,t may cause wind generators and customer induction motors totall, provoking a voltage collapse.

The results obtained in the described system, with the normalopological configuration, show that, as expected in a trans-

ission system with such a meshed structure, there were noarge changes in voltage profiles, since the transmission grid
as a considerable capacity of controlling voltages by exploit-ng OLTC transformers and synchronous generators in the mainower stations. This was verified although there was a consid-rable amount of reactive power injected by the DG plants into
itwr

sses for the six different scenarios.

he grids during the off-valley hours, which contributes to theeactive power support, but is regulated according to the needsf the system.

Variation in active losses in the studied system can bebserved in Fig. 4. These changes are described for the six dif-erent scenarios, before (2000 S) and after (2000 C) the inclusionf the DG output.

From Fig. 4, it is observed that there is a small increase inctive losses for the wet scenarios and a decrease in losses forhe dry scenarios. This results from the fact that, in this network,he large concentration of DG coincides geographically with theig hydro generation area.

However, if the increase in DG takes place in industrial areas,learly a reduction in active losses occurs with consequent largeconomic benefits that are added to the fact that thermal pro-uction is reduced in proportion to the amount of the DG powerelivered to the network.

In any network, branch loadings are increased or decreasedepending on the production level of the conventional plants. Inhe situation analysed for the Portuguese network where theres a large concentration of DG in rural areas, two different casesere detected:

Wet periods, where there is an increase in branch conges-tion levels connecting the main hydro production areas, in thenorth, and the north–south transmission corridors;Dry periods, where there is a clear reduction in branch loading,since the contribution of DG production (concentrated in thenorth) decreases the usual south–north power flows.

If DG production is concentrated in the large urban and indus-rial areas the congestion levels decrease in most of the branches,or all the conventional production scenarios, which is an impor-ant technical and economical benefit.

A complementary steady state analysis can be performed forach scenario through a probabilistic load flow or a fuzzy loadow, where the DG productions are replaced, respectively, byrobabilistic distributions or fuzzy distributions, namely whenhere is not enough statistical data do built a probabilistic model,ccording to the specific characteristics of the primary energyesources. Such an analysis will produce, instead of determin-
stic values, a set of distributions for the power flows in theransmission lines and for the voltages in the system buses. Thisill provide a complete picture about the impacts in the network
esulting from the variability of the DG resources [15].

1 System

sabcltc

6

gbsotnb

Dtbiloct

liscsbtipp

iaolSittpt

6

sst

notqiwd

orpgdaapobvbIdb

tdDss

T

w

-

-

-

wsw

6

wdsc


From this analysis, one can conclude that the impacts on theteady state operating conditions of the transmission networkre strongly dependent on the consumption levels of the distri-ution side, producing clear benefits for the grid when DG isonnected in large consumption areas and creating some prob-ems (specially overload situations) for those scenarios wherehe DG production coincides with the production from largeonventional units, situated in low load density zones.

.2. Contingency analysis

The impacts from the connection of DG on (n − 1) contin-ency analysis need also to be addressed, since the increase inranch loadings may create unexpected overload problems inome branches after a given contingency. In fact, for the usualperation strategy, the system may be considered as secure fromhe steady state point of view and become insecure after the con-ection of DG in the network, because these new injections wille responsible for some the overloads that may occur.

This issue may become the limiting factor to the increase ofG production in areas with low load consumption levels. Con-

ingency analysis can therefore be performed in a way that it cane used to identify the maximum allowable DG production levelsn export areas. For that purpose a sensitivity analysis regardingoad congestion in affected branches, relatively to the increasesf DG injection in system buses, can be performed in the post-ontingency network to identify the degree of responsibility ofhese productions in branch overloads.

When those branch overloads are detected, this means thatine up grade or construction of new additional lines are neededn these network areas. In order to assure a robust transmissionystem, the connection to large wind farms requires also, as foronventional generation plants, enough redundancy. A discus-ion on who should pay for the system reinforcements can thene developed. Either the transmission system operator pays forhose investments, recovering these costs in the tariffs, accord-ng to a general energy policy of the country related with theromotion of renewable energies, or the promoters of the DGlants share these costs amongst them in some way.

Such an impact evaluation approach needs to be developedn carefully, since, if this contingency analysis is conducted indeterministic manner for a scenario of simultaneous nominalr near nominal DG production level, it will produce seriousimitations to the integration of DG in those power export areas.uch coincidence in DG production will rarely take place, and

f this happens the right approach should be to allow genera-ion curtailment, for a limited period of time. This will enablehe connection of a large amount of DG capacity and may bearticularly suitable for wind generation, as this generation cur-ailment is likely to be required during only short periods of time.

.3. Protection schemes

Apart from own protection schemes that each generatorhould possess [11], DG plants are often required to install aet of protection systems for interconnection with the local grid,o ensure that the production plant will be disconnected from the

-

--

s Research 77 (2007) 1189–1203

etwork once a fault is detected in the grid. In Portugal, this setf protection-schemes includes three phase overcurrent relays,hree phase under and over voltage relays, under and over fre-uency relays and a zero sequence voltage relay (used to detectn an efficient way impedance earth faults in the distribution net-ork). Usually these relays are set to instantaneous operation,ue to the need to assure the safety of maintenance staff.

In a scenario of large participation of DG, concentrated inne area, a simultaneous miss operation of the interconnectionelays of these DG plants may provoke serious system operationroblems. In fact, if a short-circuit takes place in the transmissionrid and its effects on voltage are propagated downstream to theistribution level, it may provoke the disconnection of a largemount of DG production. Since these DG plants are tripped,sudden increase in power flows coming from upstream, takeslace in this network area, such that it may provoke cascadingperation of overcurrent relays in distribution and transmissionranches, as a result of overload situations. Furthermore, theoltage profiles in the area where the DG was disconnected cane seriously affected and initiate a voltage collapse phenomena.n the next section, some of the consequences of such an inci-ent are presented and analysed in terms of impacts on dynamicehaviour of the system.

This means that coordination in time delays between the sys-em protection strategy of the transmission grid and the timeelays of the voltage interconnection protections of the largerG plants (that must have a different regulation than 0 s) must be

tudied and implemented. The philosophy of such coordinationhould be:

tg < TDG < Trec

here:

Ttg—the largest fault elimination time for a short circuit in thetransmission grid.TDG—time delay of the interconnection protection of the DGplants.Trec—reclosure time in MV feeders at the MV outputs of theHV/MV substation (usually 300 ms).

However, this approach should never compromise, in anyay, the selectivity of the protections between the transmis-

ion and distribution networks and the selectivity of protectionsithin the distribution grid itself.

.4. Dynamic behaviour

The increase in the share of DG production in electrical net-orks is more and more requiring an analysis of the systemynamic behaviour of some incidents that may occur. Thesetudies should be developed carefully taking the following intoonsideration:

DG dynamic modelling according to the energy resource andadopted conversion systems;aggregation of DG generating units according to their type;definition of a set of critical disturbances.

J.A.P. Lopes et al. / Electric Power Systems Research 77 (2007) 1189–1203 1199

Ds

6

btgae

vaptltnfo

fmtcosaabtdwipotii

6

totgvc

Table 1Characterisation of disturbances

Disturbance 1 Disturbance 2

Characterisation Disconnection of one ofthe interconnection withSpain (150 MW import)

Short circuit in one 220 kVtransmission line near asubstation in the north ofPortugal, eliminated after100 ms, provoking thedisconnection of 500 MWof DG production

Scenario 1 500 MW of DGproduction for a wet peakload period

–

Scenario 2 1500 MW of DG 1500 MW of DG production

--

6

tbedTaotPua

In Fig. 6, the frequency behaviour of the system is depicted.Fig. 7 describes the dynamic behaviour of the power flow in oneof the 400 kV remaining interconnections for both DG produc-tion levels. The solid line corresponds to the system behaviour

Fig. 5. Representation of equivalent DG units.

From the transmission point of view and for this analysis, theG units should be aggregated according to generation type,

uch that they could be represented as shown in Fig. 5.

.4.1. DG modelingA proper dynamic modelling of the DG units, for dynamic

ehaviour studies, is a key issue to get an adequate idea ofhe impact in the network resulting from the presence of theseeneration units following some disturbances. In such dynamicnalysis one must remember that only low frequencies phenom-na are of interest.

Synchronous machines should be modelled through the con-entional state equations that describe the electrical machine,utomatic voltage regulator, swing equation, speed regulator andrimary machine [16]. This modelling depends however on theype of plant. Usually mini hydro plants do not have speed regu-ation. This is also often the case with cogeneration plants, wherehe speed regulator is not activated when operating in grid con-ected mode. In some industrial plants that have cogenerationacilities the speed regulator becomes active only in isolatedperation mode.

Asynchronous generators can be modelled in a proper way,or grid dynamic behaviour analysis, through a third orderodel with one mechanical equation and two electrical equa-

ions, as described in [13]. Generators with power electroniconverters can be modelled as current controlled sources [13]r active and reactive power sources [17] since the fast tran-ients in these converters, related with electronic switches oper-tion and control, are of no interest in such analysis. Theggregation of synchronous and asynchronous machines cane made adopting the usual procedures used for the aggrega-ion of conventional units in per unit [16,18], neglecting theistribution grid connections situated downstream. Generatorsith power electronic interfaces can be modelled as mentioned

n the previous paragraph, assuming an equivalent power out-ut equal to the sum of the individual outputs of each onef these units [19]. An important issue to be mentioned ishat machines with full grid electronic interfaces, character-zed by no inertia, may become a problem regarding stabilityssues.

.4.2. Critical disturbancesHaving in mind the characteristics of the DG units namely

he fact of not participating in frequency regulation, the typef relay settings imposed to the interconnection protection and
he intrinsic nature of the power electronics interface of staticenerators, that are quickly disconnected from grid when largeoltage drops occur, some of the most critical disturbances to beonsidered are:
production for a wet peakload period

for a wet peak load period

short circuit in the transmission grid;loss of an important conventional unit or loss of an importantinterconnection line.

.4.3. Dynamic behaviourIn order to illustrate the type of impacts that some dis-

urbances in the transmission grid can produce in the systemehaviour, results obtained for a generation scenario of anquivalent of the Portuguese/Spanish interconnected system areescribed next. The simulated disturbances are summarised inable 1. The DG units considered in these simulations werebout 70% of synchronous machines (with low inertia) and 30%f asynchronous ones, distributed all over the Portuguese side ofhe network, with a larger concentration in the north part of theortuguese system. The mechanical power of these generatingnits was kept constant and equal to the pre-fault values duringll the simulations.

Fig. 6. Frequency behaviour after disturbance 1.

1200 J.A.P. Lopes et al. / Electric Power Systems Research 77 (2007) 1189–1203

Fi

wrT

taHnlotbt

bc1ep

Fi

tatg

fttaldilirtT

ig. 7. Power flow behaviour after disturbance 1 in one the remaining 400 kVnterconnections.

ith 500 MW of DG connected to the grid, whilst the dashed lineepresents its behaviour for a 1500 MW DG penetration level.hese simulations were performed using the PSS/E software.

Since the DG plants do not participate in frequency regula-ion, one can see from these results that the frequency excursionsre larger for the situation where the DG penetration is larger.owever, due to the strong interconnection with the Europeanetwork, the frequency deviations are not very important in abso-ute value. It is also noticeable from the analysis of Fig. 7 thatscillations in power flow in one of the remaining interconnec-ions are larger and less damped in the scenario characterizedy 1500 MW of DG production. These behaviours result fromhe non-participation of the DG plants in frequency regulation.

In order to illustrate the consequences of the second distur-ance, Fig. 8 depicts the system frequency behaviour. In this
ase the solid line corresponds to the system behaviour with500 MW of DG, where 500 MW were lost during the faultlimination period, due to the actuation of the interconnectionrotections of the DG plants, and the dashed line corresponds to
Fig. 8. Frequency behaviour after disturbance 2.

i

dpbdt

aoamttlpnsclce

ig. 9. Power flow behaviour after disturbance 2 in one the remaining 400 kVnterconnections.

he system behaviour after a short-circuit in the same location, inscenario of only 500 MW of DG production situated far from

he short-circuit point (and therefore not disconnected from therid).

From the results described in Fig. 8, one can observe thatrequency of the system, for the scenario with larger penetra-ion of DG and disconnection of part of it, reveals, as expected,hat frequency stabilized at a different value from 50 Hz. Sincelarge amount of production in the control area of Portugal was

ost, an increase in power import took place (90% of the pro-uction lost is afterwards imported), as it can be seen from thencrease in power flow in one of the interconnection lines (solidine in Fig. 9). The lost DG production is mainly compensatednitially from the European network, since the primary reserveegulating capacity of the European system is much larger thanhe primary regulating capacity of the Portuguese control area.his requires afterwards the activation of the secondary reserves

n the Portuguese control area through the AGC operation.From the analysis of these results, one can conclude that the

ynamic behaviour of the system can be strongly affected by theresence of DG units, not only because of its specific behaviour,ut also because of the operation of their protection devices orue to the intrinsic nature of the electronic interfaces of the unitshat use this technology.

Power systems with large shares of hydro production will beble to respond in a very fast and effective way to such typef AGC requests. Systems with pumping reversible units willlso be able to deal effectively with these situations. This alsoeans that systems with reduced interconnection capabilities,

hat may become aggravated in case of large commercial energyrading between control areas, may face serious stability prob-ems in these situations, since such incidents may subsequentlyrovoke a sequence of cascading events leading to the discon-ection of the interconnections due to overload conditions. Audden reduction in wind power production, not properly fore-
asted, may also lead to overload problems in interconnectionines, which will require in the future the development of areaontrol error (ACE) monitoring tools to identify, in advance, thexpected behaviour of the system regarding such incidents. In

System

aawi

7

neDmiwtam

nttlentiutntc

castat

8

eidcirtotbclcie

bscrsadoste

8

Uor

ts2ret‘2Tytbbavt

acNwSb

8

fddr

fe


ddition, studies on the mobilisation of system load curtailmentre required, since these control measures will help in dealingith such situations and will also lead to the effective use of

nterruptability contracts.

. Development of the transmission network

Planning the expansion of a transmission network in a sce-ario of large integration of DG requires knowledge about thevolution of the load and the inclusion of estimates on how theG will grow. These estimates of DG production should beade in terms of energy and capacity to be installed, accord-

ng to the characteristics of the primary resources available, asell as their geographic distribution. This will enable the iden-

ification of conventional production needs, having in mind thevailability of the DG resources and system reserves require-ents (secondary and tertiary reserves).The development of the DG demands the availability of a

etwork to receive the DG output. This can be a difficult mat-er if the DG development happens to be in remote areas. Twoypes of uncertainties can be identified when solving this prob-em: (a) the location of the DG plants; (b) the level of expectednergy and power to be produced. The definition of a plan-ing strategy for the expansion of the transmission grid requiresherefore a certain degree of desegregation, at a regional level,n terms of the possible location of the DG production. These of geographic information systems, together with modelshat describe the resources availability and complementary eco-omic and environmental models, can be used to identifyinghe regional areas where DG production becomes attractive andould therefore appear and require connections to the grids.

Different planning alternatives need then to be considered,haracterised by investments costs and robustness indices (suchs non-delivered energy, losses, repressed demand and repres-ion on connection of DG). The transmission expansion solutiono be adopted should therefore be taken from the analysis of thettributes of these alternative solutions. A deeper discussion ofhis issue is, however, out of the scope of this paper.

. Opportunities for increasing DG penetration

Experience in the last decade has shown that in order to accel-rate the connection rate of DG, be it conventional or renewable,t is necessary to either provide incentives to renewable DGevelopers and network operators or alternatively to mandate theonnection of renewable DG under a regime of preferential feedn tariffs (the German case). The primary reason for support toenewable DG technologies is that for various legitimate reasonshey are still relatively expensive sources of energy and more-ver tend to disrupt the normal way that power systems are usedo operating. And yet they are desirable technologies that muste supported because they are good for the environment and alsoontribute to sustainability and energy security objectives. In the
ong term, there is no doubt that these technologies will becomeompetitive as the price of fossil fuels rise due to ever grow-ng demand worldwide for these fuels especially in the growingconomies in Asia (China and India). Already there is a strong
mtli

s Research 77 (2007) 1189–1203 1201

ody opinion suggesting that demand for oil is likely to outstripupply in the near future. A sustainable energy future thereforeannot be predicated on fossil fuels and must rely on new andenewable technologies. Many countries in Europe now haveome form of incentive scheme to support the uptake of renew-ble sources of energy. There is clearly a lot more that must beone to support DG in general especially from the perspectivef integrating DG in the planning and operation of transmis-ion and distribution networks. Some of the key opportunities inhis regard are discussed below drawing examples from the UKlectricity market.

.1. Government renewables incentives

Some form of renewable incentive schemes are in existence inK, Spain, Germany, Sweden, Netherlands, Norway and manyther EU countries. Most of the incentives benefit distributedenewable generation.

In April 2002, a renewable obligation was introduced inhe UK for electricity suppliers where they have to supply apecified percentage of their energy from renewable sources (in001–2002, this was 3%). This can be fulfilled by purchasingenewable obligation certificates (ROCs) from accredited gen-rators. ROCs can be sold separately from the energy. Suppliershat are unable to fulfil their obligation are required to pay abuy-out’ price to OFGEM for part or the entire Obligation. In002, the ‘buy-out’ price was set at £30/MWh until April 2003.his is adjusted by OFGEM based on the retail price index eachear. In 2002, the value of ROCs appears to be comparable tohe value of energy sold and as such, provides large financialenefits to renewable generation. The money obtained from theuy-out fund is then recycled to suppliers based on the percent-ge of ROCs purchased. Being a market based mechanism thealue of ROCs will depend on the demand for them, which isurn driven by the extent to which the target has been attained.

In Spain, income from environmental incentives comes fromcombination of renewable energy premium, subsidies and taxoncessions. A new National Energy Plan (“Plan Energeticoacional”) has been prepared and presented in September 2002,hich will guide the energy policy for the period up to year 2011.ustainability is likely to be the focus of the policy driven alsoy reduction of CO2 emissions.

.2. New security standards recognising DG contribution

A survey of electricity markets in Europe indicates that apartrom the UK the majority of countries do not have explicit well-efined security standards. Such standards are important as theyrive network investment as will become clear in the followingeview of the role of security standards in the UK.

Prior to privatisation, the Electricity Council was responsibleor setting and maintaining a range of common technical andconomic guidance documents, including Engineering Recom-
endation P2/5. ER P2/5 was intended to be used as a guide
o system planning and design. The fundamental principle out-ined within ER P2/5 is that there should be sufficient capacityn the system such that, in predefined outage situations, cus-

1 System

tapmcntrtCtbptid

dTTalcp

cgiqotfgPp

ctsgsatsaia4

iDi

8

w

Detptdfirdmbtn

8c

noiirsbe

otscr

9

cp(usottatotwpeg


omers continue to receive a supply or have it restored withinn acceptable time period. ER P2/5 defines the network designhilosophy and, in the context of network pricing, the require-ents to comply with the security standard are a key network

ost driver. Historically, the structure of electricity distributionetworks was driven by an overall design philosophy developedo support large-scale generation technologies. The level of secu-ity in distribution networks is defined in terms of the time takeno restore power supplies following a predefined set of outages.onsistent with this concept, security levels on distribution sys-

ems are graded according to the total amount of power that cane lost. In general, networks have been specified according to arinciple that the greater the amount of power which can be lost,he shorter the recommended restoration time. This philosophys formalised in the ER P2/5, and this standard is a part of theistribution network licence conditions.

ER P2/5 contains two tables. Table 1 states the minimumemand that must be met after certain specified circuit outages.his level is dependent on the group demand or class of supply.he amount of demand that can be supplied depends upon thevailable circuit capacities and critically, the contributions fromocal distributed generators. Table 2 in ER P2/5 specifies theontribution that can be attributed to generation connected to aarticular load group.

Given that ER P2/5 was developed in 1970s, it does noturrently recognise many of the modern forms of distributedeneration presently being connected into distribution networksn the pursuit of Government climate change targets. Conse-uently, it is not possible to recognise the security contributionsf many new forms of distributed generation. It is anticipatedhat ER P2/6 will supersede ER P2/5 in the near future. This willacilitate the inclusion of security contributions from distributedeneration during network planning [20]. In other words, the ER2/6 will specify the critical condition for network design in theresence of distributed generation.

The new ER P2/6 will provide a basis for quantifying theontribution that DG makes to system security, i.e. the extento which DG can reduce the demand for network facilities andubstitute for network assets. In the network planning stage,eneration contribution to network security, as specified by thetandards, can be interpreted as firm generation output availablet peak demand (during specified period of time for intermit-ent generation). For various generation technologies, this ispecified through a series of contribution factors that take intoccount the number of generators, their availability and operat-ng regimes. For example, a CHP generator (spark ignition) withn availability of 60% can contribute to network security with0% of its installed capacity.

This is a pre-requisite for establishing cost reflective pricingn networks with DG that can recognise the positive impact thatG may have on network expenditure (reduction of network

nvestment cost) and that can reward generators accordingly.

.3. Incentives to DNOs to connect DG

It has been recognised for sometime now that distribution net-ork companies generally do not have an incentive to connect

atha

s Research 77 (2007) 1189–1203

G. In practice DG is in fact viewed as a nuisance disruptingstablished operating and planning practices with little gain forhe network operators. Moreover in the UK, for example, afteraying an initial connection charge (based on the deep connec-ion charge policy), generators are exempt from paying ongoingistribution use of system (DUoS) charges. This scenario bene-ts neither the network operator nor DG. DG, for example is notecognised for the positive impact it has on losses, investmenteferral and better utilisation of existing assets. In acknowledg-ent of this unsatisfactory situation in the UK, a framework has

een created within which network companies are incentivisedo connect distributed generators based on capacity of DG con-ected and level of DG utilisation.

.4. Review of methodology for distribution Use of systemharges

As indicated above the current regime for DUoS charges doesot deal adequately with the presence of DG. The UK Regulatorf Gas and Electricity Markets (Ofgem) is in the process review-ng the methodology applied to calculate DUoS charges. Theres a desire to develop a DUoS calculation methodology that willecognise the presence of DG and its contribution to systemecurity as well as to losses. The methodology is likely to beased on shallow connection charge policy and have economicfficiency as its key driver.

This review clearly is an opportunity to improve the viabilityf DG by treating it in a fair and consistent manner recognisinghe costs/benefits that DG brings to the system within the use ofystem pricing structure. The approach being taken in the UKan be used by other countries in the EU and elsewhere in theeview of their own distribution network pricing methodologies.

. Conclusions

This paper has presented an overview of the key issues con-erning the integration of distributed generation into electricower systems that are of most interest to the stakeholderspower system planners and operators, policy makers and reg-lators, DG developers and customers) in the electrical energyupply industry today. A detailed discussion has been presentedf the main drivers of DG integration in many countries aroundhe world. These issues are still just as important and compellingoday, perhaps even more so, as they were more than a decadego when DG became recognised as an important issue in elec-ricity supply. An overview of the main challenges that must bevercome in the integration of DG into electricity supply sys-ems has also been presented. In this paper, particular emphasisas placed on the need to shift network planning and operatingolicies away from the fit and forget policy of connecting DG tolectric power systems to a new more appropriate policy of inte-rating DG into power system planning and operation though
ctive management of distribution networks. Some of the oppor-unities that could be exploited in support of the integration andence greater penetration of DG into electric power systems arelso explored.

System

tptapcatac

mlstbttdp

-

-

--

-

tgpfvsr

ar

histmtng

rsiatm

R

[

[

[

[

[

[

[[

[

[19] J.G. Slootweg, W. Kling, Impacts of distributed generation on power system


The discussion of the impacts of large DG integration inhe transmission network shows that there are no dramaticroblems for system operation. However, as the DG pene-ration rises, changes in the operation policy of the networkre needed in order to reduce the magnitude of some newroblems. The needed operational changes are in the areas ofoordination of protection systems and coordination of oper-tion regarding ancillary services. Investments in communica-ion infrastructures and in development of new tools for EMSnd DMS environments are needed to support the requiredhanges.

In areas where the DG concentration is large it is recom-ended to allow generation curtailment, in general for only

imited periods of time, in order to comply, when neces-ary, with the transmission system security limits of opera-ion. However, management of DG unit capabilities shoulde used also to help manage the local distribution grids andhe production/transmission system through the already men-ioned additional functionalities of DMS or through the intro-uction of local production dispatch control centres, able torovide:

capability of aggregation of hourly DG production levels, inorder to provide power forecasts;ability to limit production injections in the transmission gridthrough local DG control;ability to limit DG production ramping rates, if necessary;management of reactive power support (presently alreadytechnically feasible);management of active power reserves, thus enabling reduc-tions in secondary reserves (which can be easily implementedin the present technology scenario).

As far as wind power is concerned and when production cur-ailment is foreseen, it will be necessary to develop combinedeneration/storage solutions, where the electrical wind powerroduction curtailed for the network could be stored (producingor instance hydrogen or pumping water to upper level reser-oirs) and later injected in the system during periods of low windpeeds. Naturally, such capabilities need to be remunerated aseserve services.

Also grid codes need to be revised, having in mind the oper-tional changes that the introduction of large shares of DG willequire in the system.

If located in large industrial and residential areas, DGas significant benefits as it decreases the need for build-ng new lines and installing new transformers. In rural areaserved by weak grids the increase in DG will require addi-ional investments in transmission infrastructures. These invest-

ents have to be decided carefully, having in mind theype of line scheduling that may result from the intermittentature of the power produced by these units (especially windeneration).

[

s Research 77 (2007) 1189–1203 1203

It is also important to mention that the development of DG inemote areas cannot by made on the basis of firm transmissionervices, where the generating plants must reserve transmissionn advance and pay for it independently of the use that is made,s it happens in some systems. Such an approach undermineshe development of DG, especially the plants that exploit inter-

ittent renewable primary energy sources.

eferences

[1] CIRED Working Group WG04, Dispersed Generation, June 1999.[2] CIGRE Working Group WG 37-23, Impact of increasing contribution of

dispersed generation on the power system, 1997.[3] COM (97) 559, Communication from the European Commission: energy

for the future, renewable sources of energy, White Paper for a CommunityStrategy and Action Plan.

[4] European Commission, Directive of the European Parliament and the Coun-cil on the Promotion of Electricity from Renewable Energy Sources in theInternal Electricity Market, 2000.

[5] N. Hatziargyriou, A. Zervos, Wind power development in Europe, Proc.IEEE 89 (12) (2001) 1765–1782.

[6] Wind Directions, vol. XXII, No. 1, November 2002.[7] Ofgem, Electricity distribution price control review Appendix—Further

details on the incentive schemes for distributed generation, innova-tion funding and registered power zones, June 2004 http://www.ofgem.gov.uk/ofgem/work/index.jsp?section=/areasofwork/distpricecontrol.

[8] N.D. Hatziargyriou, T.S. Karakatsanis, M. Papadopoulos, Probabilisticload flow in distribution systems containing wind power generation, IEEETrans. Power Syst. 8 (1) (1993) 159–165.

[9] N. Hatziargyriou, T. Karakatsanis, G. Strbac, Connection criteria forrenewable generation based on probabilistic analysis, in: Sixth Interna-tional Conference on Probabilistic Methods Applied to Power Systems,PMAPS’2000, Funchal, Madeira, Portugal, September 25–28, 2000.

10] EC Contract JOR3-CT98-0201, Final Report, Electricity Tariffs andEmbedded Renewable Generation, July 2000.

11] N. Jenkins, R. Allan, P. Crossley, D. Kirschen, G. Strbac, Embedded Gen-eration, IEE Power and Energy Series 31, London, 2000.

12] Ilex Energy Consulting with the Manchester Centre for ElectricalEnergy, UMIST, A report for DTI on Ancillary Service Provisionfrom Distributed Generation, September 2004, URL: http://www.dti.gov.uk/renewables/publications pdfs/dgcg000300000.pdf.

13] J. Pecas Lopes, Integration of dispersed generation on distributionnetworks—impact studies, in: Proceedings of the IEEE Winter Meeting,N.Y., February 2002.

14] P. Smith, H.K. Clark, Impact of increasing wind generation on the trans-mission system in the Republic of Ireland, in: Proceedings of the CIGRESymposium Neptune 1997, Romania, 1997.

15] T.P. Leao, M.A. Matos, Distribution Planning with Fuzzy Loads and Inde-pendent Generation, in: Proceedings of the 14th International Conferenceand Exhibition on Electric Distribution, 1997.

16] P. Kundur, Power System Stability and Control, McGraw-Hill, 1993.17] J.G. Slootweg, S. de Haan, H. Polinder, W. Kling, Modeling new genera-

tion and storage technologies in power system dynamics simulations, in:Proceedings IEEE Summer Meeting, Chicago, July 2002.

18] J. Machowski, J. Bialek, J. Bumby, Power Systems Dynamics and Stability,John Wiley & Sons, 1997.

transient stability, in: Proceedings of the IEEE Summer Meeting, Chicago,July 2002.

20] R. Allan, G. Strbac, P. Djapic, K. Jarrett, Methodology for developing ERP2/6, DTI Report, April 2004.

http://www.ofgem.gov.uk/ofgem/work/index.jsp%3Fsection=/areasofwork/distpricecontrol

http://www.dti.gov.uk/renewables/publications_pdfs/dgcg000300000.pdf

Integrating distributed generation into electric power ...distributedwind.org/assets/2013/05/Integrating-distributed... · Integrating distributed generation into electric power ...

Documents