Quantifying the Impact of Connection Policy on Distributed Generation

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 22, NO. 1, MARCH 2007 189

Quantifying the Impact of Connection Policy onDistributed Generation

Andrew Keane, Student Member, IEEE, Eleanor Denny, Student Member, IEEE,and Mark O’Malley, Senior Member, IEEE

Abstract—Increasing connections for distributed generation(DG), and in particular, for wind generation, are being soughtin power systems across the world. These increased applicationspresent a significant challenge to the existing connection policiesof distribution network operators. In particular, nonfirm access tothe network has been proposed as a method to increase the pene-tration of DG. The impact of the connection policies arising fromnonfirm access are investigated in detail here. The Irish system isused as a case study, and with reference to the available energyresource and network parameters, the costs and benefits of DG aredetermined under a number of planning policies. The costs andbenefits assessed include connection and cycling costs along withemissions, capacity value, and fuel bill saving. It is shown that asignificant increase in the net benefits of DG is gained if the appro-priate connection policy is utilized from the outset, and conversely,significant costs are incurred if ad hoc policies are employed. Fur-thermore, it is shown that nonfirm access has the scope to facilitatea significant extra amount of DG capacity.

Index Terms—Costs, dispersed storage and generation, en-ergy resources, linear programming, power distribution planning,power system economics, wind power generation.

I. INTRODUCTION

D ISTRIBUTED generation (DG) can be defined as small-scale generation, which is not directly connected to the

transmission system and is not centrally dispatched. The con-nection of DG has led to a change in the characteristics of thenetwork. If increasing levels of generation are to be accommo-dated, then there should be a change in thinking regarding theplanning and design of the distribution network. Wind genera-tion is the fastest growing form of DG with significant penetra-tions connected already in many countries [1].

The traditional approach to DG planning, on a first come firstserved basis, is becoming outdated as the volume of applica-tions increase. Indeed, in some countries it has already beenreplaced [2]. Another traditional approach to DG planning hasbeen to permit access to the distribution network only on a firmbasis. The amount of firm access granted under the connec-

Manuscript received July 12, 2006; revised October 23, 2006. This workwas conducted in the Electricity Research Centre, University College Dublin,which was supported in part by ESB Networks, in part by ESB Powergen, inpart by ESB National Grid, in part by Cylon, in part by Viridian, in part bythe Commission for Energy Regulation, and in part by Airtricity. The work ofA. Keane was supported by Sustainable Energy Ireland through a postgraduateresearch scholarship from the Irish Research Council for Science Engineeringand Technology. Paper no. TEC-00315-2006.

The authors are with the School of Electrical, Electronic and MechanicalEngineering, University College Dublin, Dublin 4, Ireland (e-mail: [email protected]; [email protected]; [email protected])

Digital Object Identifier 10.1109/TEC.2006.889618

tion agreement to a distributed generator is the level of outputat which they can always operate without violating any of theN–1 constraints on the network. The impact of these techni-cal constraints on the penetration of DG has been demonstratedin [3] [4], along with the network sterilization effect, which canoccur if DG capacity is not allocated with full account of theseconstraints [5], [6].

Nonfirm access refers to output greater than the firm amount,at which generators may be allowed operate dependent on theload and generation levels throughout the year. Nonfirm accessutilizes some form of active control to manage the constraintbreaches. Previous work has shown the scope for the optimiza-tion of plant mix for the maximization of energy [7]. There area number of costs and benefits associated with the various typesof DG. Previous work has attempted to quantify the net benefitsof DG [8], where a number of benefits such as reduced lossesand voltage profile were assessed.

This paper seeks to assess the impact of high penetration ofDG, taking into account the likely DG plant mix. It does notattempt to place an upper limit on the amount of DG that can befeasibly connected to an entire distribution network. The impactof DG on the transmission system is not included here. The dis-patch model used is a single busbar system, and the aim of thispaper is to assess the impact of DG on the distribution system.In particular, the voltage constraint is the dominant constraintand can be managed on a nonfirm basis. The implementation ofnonfirm management of constraints requires a change in the tra-ditional connection policy of network operators. Furthermore, itleads to a change in the optimal allocation of DG as calculatedusing previously established methodologies [4]–[6]. It is issuessuch as these, arising out of the change in connection policy,that lead to a change in the costs and benefits of DG. Thesecosts and benefits are quantified here to assess the impact ofconnection policy on DG, and more specifically, to assess theimpact of nonfirm access.

The Irish system is taken as a case study. Drawing on es-timates of the Irish renewable energy resource [9] and systemdata [10] [11], the impact of different connection policies forDG is assessed, quantifying in each case the costs and bene-fits that result. The importance of a clear set of objectives atthe start of the connection process is highlighted, and it willbe demonstrated that significant costs can be incurred throughthe employment of ad hoc methodologies. The optimal firm al-location of the available energy resources may not lead to theoptimal nonfirm allocation at a later stage. This paper addressesthe question of which policy is best for facilitating higher pen-etrations of DG. Different individual allocations impact on the

0885-8969/$25.00 © 2007 IEEE

190 IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 22, NO. 1, MARCH 2007

costs and benefits of DG, and it is these costs and benefits thatare quantified here under a number of connection policies.

Section II describes the methodology used to determine theimpact of various connection policies. Section III describes testsystems and energy sources. Results and discussion are given inSection IV illustrating the impact that the connection policy hason the cost and benefits of DG. Comparisons between differentcases are given. Conclusions are given in Section V.

II. METHODOLOGY

The optimization of access to the distribution network hasbeen demonstrated for a single section of distribution networkin [5] and [7]. Here, the integration of DG across the wholesystem is investigated. The distribution network is optimizedusing previously established methodologies. These methods areapplied under three different connection policies, and the costsand benefits are then quantified to determine the impact of con-nection policy on DG, and more specifically, on wind generation.

A. Objective Function

The objective of the methodology is to maximize the amountof DG energy per euro of investment by making the best use ofthe existing network assets and available energy resource. Thisis done subject to the technical constraints on the network. Theoptimization problem is formulated as a linear program, withthe amount of constraint breaches that arise with nonfirm accesstaken into account. The load factors of each energy resource areincluded, meaning that the available capacity is allocated basedon the amount of energy that is delivered. The objective functionis given in as follows:

J =M∑

j=1

N∑

i=1

PAvail jPlantijLFj

ConnCostijνi(1)

where PAvail j is the jth available energy resource. Plantij arethe control variables representing the fraction of PAvail j allo-cated to the ith bus. LFj is the load factor of the jth energyresource. νi gives the total voltage sensitivity of the ith bus topower injections at all other buses. ConnCostij is the connec-tion costs of the jth energy resource at the ith bus. M and N arethe number of energy resources and buses respectively. The ge-ographical dispersion of DG impacts the connection costs. Theconnection costs of each resource to each bus are determinedand are minimized in the objective function. The objective func-tion is maximized with respect to the technical constraints in(3)–(6), given in the Appendix [5]. These constraints limit theDG allocation based on the technical characteristics of the net-work.

This objective function in (1) maximizes the energy harvest-ing capability of the existing network. It maximizes it withrespect to the technical constraints. The available capacity is avaluable asset and should be optimally utilized. This objectivefunction maximizes the energy harvested from the available ca-pacity by using the load factors of the generation. Load factorsexpress the energy output of a generator as a fraction of the max-imum possible energy output that is produced by a generator in

a year. In addition, it is now assumed that nonfirm manage-ment of the voltage constraint is permitted, thus allowing anincrease in the permissable capacity. This nonfirm capacity isoptimized through the use of bus voltage sensitivities νi. Thesesensitivities are included in the objective function as a meansof reducing the occurrence of overvoltage conditions, therebyreducing the amount of energy curtailed and enabling easiercongestion management on the distribution network.

B. Nonfirm Constraint Management

The local network constraints have previously been analyzedin detail in [3]. The voltage constraint, given in (7) in Appendixis the key constraint in the analysis. Voltage rise is one of thedominant technical constraints on DG. The assessment of thevoltage constraint at a N–1 peaking condition can present a sig-nificant barrier to further penetration of DG. While infrequent,it is important for the operation of the system that the voltagestays within its limits. A number of voltage control techniqueshave been proposed to mitigate the voltage rise effect [12], [13].These techniques along with others facilitate non firm manage-ment of the voltage constraint. The short circuit level (SCL)may be dominant in more urban areas. However, active man-agement of fault levels is some way off and is likely to be veryexpensive [14]. Hence, it is nonfirm management of the voltageconstraint that is considered here.

C. Network Analysis

A number of representative samples of the distribution net-work were analyzed. These sections were chosen based on theirgeographical location and also on their characteristics (i.e., ruralor semi urban). Likely energy resource portfolios were associ-ated with each section. Drawing on the methods developed in [5]and [7], each section was analyzed and the optimal allocationof the energy resources was determined for three cases:

� firm;� nonfirm;� firm + nonfirm.The first case is the base case of the maximum firm allocation

optimized on a megawatt hour per € basis, i.e., the amount ofenergy per euro of connection costs is maximized. The voltageconstraint in (7) is included, and the voltage sensitivities νi areremoved from the objective function. The second case refersto the maximum nonfirm access permitted, where the voltagesensitivities νi are utilized in the objective function to reducethe instances of overvoltage. The third case examined is thecase where firm access is initially permitted, and then, nonfirmaccess is permitted at a later stage. These two nonfirm cases areoptimized on a megawatt hour per € kilovolt basis as shownin (1).

The implementation and stages of the methodology are shownin Fig. 1. It is seen that based on the technical constraints,the allocations under the two objectives are calculated. Thefirm allocation (MWh/€) is fed as a constraint into the nonfirmallocation, which yields the firm + nonfirm allocation. Eachof the three allocations are then simulated on the distributionnetwork over a year using load flow analysis. From these load

KEANE et al.: QUANTIFYING IMPACT OF CONNECTION POLICY ON DISTRIBUTED GENERATION 191

Fig. 1. Methodology.

flow analysis, the instances of curtailment are determined andused to appropriately manipulate the profiles used in the dispatchmodel. Based on the analysis of the Irish-distribution networkand energy resource, the allocations are scaled to reflect boththe likely DG penetrations of each resource and the overallcharacteristics of the Irish-distribution network. The dispatchmodel is run based on these scaled allocations for the sameyear, and the costs and benefits of each of the three connectionpolicies are calculated based on the results of this.

D. Annual Simulations

1) Load Flow Simulation: The allocations of DG capacityon the system are determined, and the annual energy outputfor each DG resource is ascertained. For the firm case, theannual energy output is determined using actual historical dataseries from [15]. For the nonfirm cases, over the year, constraintbreaches will occur at times of high generation and low load.Annual simulations are carried out for the network sections todetermine the time and magnitude of energy curtailment. Thesesimulations consist of load-flow calculations carried out for halfhourly data. The data included in the simulations includes activeand reactive load and generation profiles for each of the energy

resources, along with data on frequency of N–1 outages and thesending voltage at the transmission station.

When constraint breaches occur due to overvoltage, the levelof curtailment is calculated using the formula shown in (2)

PCurtail i =Vi − VMax

2νii∀N (2)

where PCurtail i (megawatt hour) is the amount of energy thatis curtailed over a half hour period at the ith bus. Vi and VMax

(kilovolt) are the voltage at the ith bus and the maximum per-missable voltage, respectively. νi is the voltage sensitivity (kilo-volt per megawatt) of the ith bus, and N is the number ofdistribution buses. νi is calculated for each bus drawing onthe analysis used in [5]. This curtailment method results in theleast amount of energy being curtailed by curtailing the gen-erator, which has the highest voltage sensitivity at that bus.The annual load flow simulation yields the amount of energycurtailed over the year, which in turn gives modified genera-tion profiles that can then be used in the economic dispatchmodel.

2) Dispatch Model: A dispatch model of the full-Irish sys-tem was developed that generates least-cost dispatches usinga linear programming formulation, which cooptimizes genera-tors’ operating and reserve levels on a half hourly basis (dis-cussed in detail in [16] and [17]). Given that renewable energyresources must receive priority dispatch in Ireland [18], it isassumed that the DG resources bid zero into the model and aredispatched according to their actual energy output. It is assumedthat the DG resources do not provide reserve. The conventionalgenerators are then dispatched according to their bid prices forenergy and reserve on a least-cost basis.

The dispatch model was run for each half hour for one year,and the generation and reserve operating levels for each gener-ator were determined, given the DG-installed capacities in eachof the three connection policy scenarios. As described above, theprofiles are then manipulated to curtail the DG output at timesof high output and low load, which correspond to the times ofcurtailment recorded from the annual load flow simulation. Thechosen test year is 2007 with the assumed load and conventionalplant mix given in [19]. These dispatches were then used to cal-culate the costs and benefits of the three different connectionpolicies.

E. Costs and Benefits

The capital cost of wind generation is assumed to be€900 000/MW with an operation and maintenance cost of€45 000/MW per year with a term of 15 years and a discountrate of 7.5% [20]. The connection costs are variable per km ofline and are taken to be €50 000/km.

An increase in the penetration of variable wind generation on asystem results in an increase in uncertainty on the electricity sys-tem as wind generation is relatively unpredictable and nondis-patchable. This results in a requirement to carry additional re-serve capacity in order to maintain system security [21], and is acost imposed on the system by increased wind generation. TheDG technology that can be considered to be dispatchable, such


as biomass, does not impose an additional reserve requirementon the system. The additional reserve requirement for the in-stalled wind generation under the three connection policies wasbased on [22].

Conventional thermal units are designed to be at their mostefficient when online and are running at a stable load. In gen-eral, units are optimized for continuous rather than cyclicaloperation, and when operating in their normal range can op-erate for relatively long periods with relatively low risk andloss of equipment life [23]. An increase in the cycling of con-ventional units, as a result of an increase in variable genera-tion on the system, can result in increased wear and tear onthe components of the machine and result in a shortening ofthe life span of the unit [23]. Cycling costs are high and canrange from €200 to €500 000 (including fuel cost) per singleON-OFF cycle depending on the type of unit [23]. Thus, anincrease in cycling as a result of an increase in the penetra-tion of variable generation on the system can have a signifi-cant impact on system cycling costs and are therefore includedhere.

Despite being variable, wind generation and the other DGresources have a capacity benefit. The extent to which a tech-nology can substitute for conventional generation is given bythe capacity value of that technology, and for nondispatchabletechnologies, the capacity value decreases with increasing in-stalled capacity [24]. An increase in the penetration of DG re-sults in a reduced requirement for new conventional generationdevelopment. Thus, the capacity benefit of DG can be deemedto be the saved cost of building and maintaining conventionalthermal generation in its place. The capacity credit for Irelandranges from 41% at low-wind penetrations to 20% at high-wind penetrations [24]. It is assumed that any new conventionalgeneration built in Ireland will be gas fired with a capital costof €687 000 per MW installed and operation and maintenancecosts of €50 000/MW per year [24], [25].

The operation of renewable generation can result in a re-duction in the operation of some of the thermal units on thesystem. The carbon dioxide (CO2) and sulphur dioxide (SO2)emissions of a thermal unit depend on the carbon and sulphurcontent of the fuel, respectively, and the quantity of fuel burnt.Thus, a reduction in the output of a thermal unit, as a resultof an increase in DG generation, can lead to a reduction inCO2 and SO2 emissions [16]. This is a significant benefit ofincreased DG penetration. Nitrogen Oxide (NOx) emissionsare more complicated and are effected by the flame tempera-ture, residence time, oxygen concentration etc. Thus, it is notnecessarily the case that a reduction in output leads to a reduc-tion in NOx emissions, and an increase in variable generationcould in fact result in an increase in system NOx emissions[16].

A reduction in the output of thermal units on the system canresult in a reduction in the quantity of fuel burnt. Thus, an in-crease in DG can result in saving the system fuel. This is asignificant benefit of increased DG generation, and in partic-ular, for wind generation that has a zero-fuel cost. All of theabove costs and benefits were calculated for each of the threeconnection policy allocations [26].

Fig. 2. Future per-county distribution of installed wind power capacity inpercentage.

III. TEST SYSTEMS AND ENERGY RESOURCE

A. Test Systems

In Ireland, DG is typically connected at 38 kV or by directfeed to the 38 kV/MV station on the MV network (10 kV or20 kV). Voltage rise tends to be the dominant constraint due tothe network in Ireland, which outside the main cities is typicallya weak rural network with a large amount of conductor. A weaknetwork means a network with a low SCL or fault level. Irelandhas a rich wind resource, especially, along its western coastline.In addition to the wind resource, there is limited scope for small-scale hydro and landfill gas along with significant amounts ofbiomass in the form of industry and forest residue [9]. Drawingon estimates of the Irish renewable energy resource [9] and [27],and its geographical dispersion are used to assess the successfulintegration of DG on a system-wide basis.

Five test systems are chosen as a representative sample of thetypes of network encountered on the Irish distribution networkwhere generation will seek to connect. Load flow and shortcircuit analysis is used to determine and formulate the variousconstraints and factors used in the optimization. Load valuesand network parameters for each section were obtained fromESB National Grid [10] and ESB Networks [11]. Fig. 2 showsthe likely future per county distribution of installed wind powerper county, reproduced from [22]. The five network sectionswere selected from along the western coastline and also in thesoutheast of the country. It is evident from Fig. 2 that it is inthese locations that connections of DG, and in particular, windgeneration will be and are being sought.

The network sections were also selected based on theircharacteristics and structure from network data available [11].They represent the types of network encountered on the Irish


TABLE IASSUMED ENERGY RESOURCES (MW)

TABLE IIALLOCATIONS UNDER THE THREE DIFFERENT CONNECTION POLICIES FOR THE

FIVE NETWORK SECTIONS (MW)

distribution network from voltage-constrained rural networks tomore urban networks, where voltage is less of a problem andSCL can become significant.

B. Energy Resources

A representative energy resource portfolio is drawn up foreach network section based on its geographical location andwith reference to Fig. 2 and [9]. Using typical values for theload factor of each type of generation [28], these resourcesare incorporated into the objective function (1). Each networksection has its own distinct assumed energy resource. Table Ishows the total assumed energy resource and load factors acrossthe five representative network sections.

IV. RESULTS AND DISCUSSIONS

A. Energy Allocations

The allocations for the five network sections were determinedunder the three connection policies using the methodology de-scribed in Section II, and are shown in Table II. Given theirsignificantly higher load factor, any available biomass and land-fill gas (LFG) are allocated the available firm capacity first. Asa result, there is very little difference, between their firm andnonfirm capacities, leading to the conclusion that when a max-imization of energy strategy is followed, forms of generationwith higher load factors are largely independent of the connec-tion policy, as they will always be connected first. The small-

TABLE IIISCALED ALLOCATIONS UNDER THE THREE DIFFERENT CONNECTION POLICIES

FOR THE WHOLE SYSTEM (MW)

scale hydro generation also has very little difference betweenits firm and non-firm allocation, but this is due to the limitedresource available. When these factors are combined with thehigh-wind resource, it is evident that it is the costs and benefitsof wind generation that are affected by the connection policy,and it is wind generation that is the focus of the results in thispaper.

Initially, it is assumed that there is no generation previouslyconnected. For the firm + nonfirm case, it is assumed that themaximum firm capacity has been connected. The nonfirm ca-pacity is then allocated on top of the existing firm capacity.Two out of the five network sections analyzed were not voltageconstrained. These network sections were located in less ruralareas, i.e., with higher load and SCLs. It was found that the SCLwas the binding constraint in these cases, and therefore, nonfirmvoltage access did not yield any further generation capacity. Theother three network sections were all voltage constrained andhad a rich wind resource, as would be typical all along the west-ern coastline of Ireland. In these cases, nonfirm access yieldeda significant amount of extra capacity. However, for one of thesections no difference was found between the two nonfirm con-nection policies. In this case, it was found that the constraintof the existing firm access did not affect the optimal non firmallocation. Although in the other two-voltage constrained net-work sections, the existing firm allocation did affect the nonfirmallocation, resulting in higher connection costs and more cur-tailment of energy.

From Table II it is evident that given a large wind resource,nonfirm access has the potential to facilitate much higher pen-etrations than firm access. Nonetheless, the question arises ofwhat is the optimal way of providing non firm access. The allo-cations shown in Table II are only for five network sections.

To quantify fully the effect of the connection policy of DGon the whole system, these allocations were scaled up based onthe total Irish distribution network and the likely penetrationsand locations of renewable energy [9], [11], [24], and [27].The scaled allocations for DG under the three connection poli-cies were determined to facilitate the quantification of the costsand benefits of DG on a system-wide basis. The installed ca-pacities and the annual energy outputs, based on actual out-put data series [15] and the total curtailed energy, are given inTable III.


Fig. 3. Output of DG generators as a percentage of their installed capacity ona sample day.

TABLE IVCONNECTION COSTS FOR WIND GENERATION

Fig. 3 shows the output profile for each of the four allocatedDG types over a single day in March on a single section ofnetwork. This day is illustrated as it is an example of a day ofparticularly high wind generation when the output had to becurtailed in the nonfirm case.

B. Wind Generation Development Costs

The total connection costs for wind under the three connec-tion policies across the whole system are given in Table IV. Aconsiderable saving in connection costs is made when a nonfirmconnection policy is pursued from the outset. In some cases, theextra nonfirm capacity can share the existing connection line,but it is seen from the results that on other occasions new con-nection lines must be built.

C. System Costs and Benefits

The DG output levels for each technology under the threeconnection policy scenarios for each day of the sample year isdetermined from the DG generation profiles from [15]. Theseoperating levels were then bid into the dispatch model with abid price of zero, ensuring priority dispatch. The resulting dis-patches were then analyzed to determine the costs and benefitsof each of the three connection policy approaches.

The additional reserve requirement for the system with in-creased wind penetration is taken from [22] and the cost in-curred is valued at €1.15/MW per hour [29]. However, giventhat the installed capacity is the same in both nonfirm cases, the

TABLE VADDITIONAL CYCLING COSTS WITH WIND GENERATION

TABLE VIEMISSIONS SAVINGS (IN TONS) WITH WIND GENERATION

TABLE VIIFUEL SAVINGS IN PETAJOULES WITH WIND GENERATION

additional reserve costs are unaffected by the connection policyand are the same in both cases.

In order to calculate the cycling cost associated with each ofthe three allocations, all of the starts and ramping excursionsover the year were counted and were converted into equivalenthot starts [23]. The cycling cost of each hot start was thencalculated according to the type of generator in question. Theadditional cycling costs over zero-installed wind capacity, undereach of the connection policies is given in Table V. It can beseen that significant extra cycling costs are incurred as morewind generation is installed on the system.

The dispatch model was run for each day of the test yearand the resulting dispatches were then used to calculate theCO2, SO2, and NOx emissions for each of the conventionalgenerators [16]. Since the wind generation displaced mostlymarginal oil-fired plants and gas plants it was seen to give asaving across each of the three emissions. Table VI below showsthe emission savings across the year compared to a no wind case.The carbon dioxide emissions are valued at €30/ton, and the SO2

and NOx emissions at €150 and €3000/ton, respectively [30].Given the magnitude of the CO2 emissions when compared tothe other emissions, the price of carbon dictates the benefitsaccruing from increased wind penetration. An emissions factor,which accounted for the price of the emissions was included inthe bid price of each generator in the dispatch model.

The additional wind generation on the system also providesa saving in fuel consumption as thermal units are displaced.Table VII shows the fuel savings with increased wind generation.The fuel prices given in Table VII were also used in the dispatchof the generators [26].


TABLE VIIINET BENEFITS OF WIND GENERATION (IN €M)

D. Net Benefits

The costs and benefits above were combined and the netbenefits of each connection policy were determined. It was foundthat although the benefits outweighed the costs under all threeconnection policies there was a significant difference in thenet benefits. Table VIII shows the net benefits of each of theconnection policies.

As is evident from Table VIII, while the two nonfirm policiesresult in the same installed capacity of generation, the costs andbenefits of wind generation show significant differences. Thereis a significant increase of 12% in net benefits under nonfirmconnection than under a firm, and then, the nonfirm scenario forthe same level of installed wind. The increases in net benefitsresults from the different pattern on connected generation acrossthe buses, which leads to altered connection costs and curtailedenergy and in turn impacts upon a number of the costs andbenefits as detailed above.

It is clear that the connection policy affects the energy outputand connection costs and furthermore that this has a consid-erable knock on effect on system costs and benefits such ascycling costs, emissions and fuel use. It has also been shownthat as some costs, such as the additional reserve cost and thecapital cost, are dependent on the capacity of the plant in-stalled, they are unaffected by the non firm connection policyemployed.

V. CONCLUSION

The impact of connection policy on DG for a number ofconnection policies has been quantified. The utilization ofthe appropriate connection policy from the outset has beenshown. There is a considerable effect on the costs and ben-efits of DG, which has been determined. It has been shownthat an equal amount of DG capacity connected can have sig-nificant differences in costs and benefits, dependent on theconnection policy. The advantages of a clear objective at theoutset of the DG planning process has been clearly shown.In addition, the significantly higher penetration of DG facili-tated by nonfirm access to the distribution network has beendemonstrated.

APPENDIX

A. Thermal Constraint

Ii < IRatedi i∀N. (3)

Here, Ii is the current flowing from generator i to bus i, IRatedi is

the maximum rated current for the line between each generatorand its corresponding bus.

B. SCL

N∑

j=1

δjTxPDG j + αTx ≤ SCLRated. (4)

Here, δjTx is the dependency of the SCL at the transmissionstation to power injections at bus j. αTx is the initial SCL at thetransmission bus with no generation present and SCLRated isthe maximum permissable SCL as laid down in the distributioncode.C. Short-Circuit Ratio

PDG i − 0.1 cos(φ)N∑

j=1

δjiPDG j ≤ 0.1 cos(φ)αi i∀N.

(5)Here, cos(φ) is the power factor at the generator.D. Transformer Rating

N∑

i=1

(PDG i − PLD i) ≤ PTrafoCap. (6)

Here, PTrafoCap refers to the rating of the transformer, and PLD i

is the minimum load level at the ith bus.E. Voltage Rise

N∑

j=1

(µijPDG ij+νijPLD ij) + βi ≤ Vmax i i∀N. (7)

Here, µji and νji refer to the dependency of the voltage levelat bus i on power injections and load at bus j, respectively, βi

refers to the initial voltage level at the ith bus with no generation.

ACKNOWLEDGMENT

The authors would like to thank ESB National Grid and ESBNetworks for the data provided, and their colleagues in theElectricity Research Centre for their help.

REFERENCES

[1] European Wind Energy Association. (2005). A blueprint to achieve 12%of the world’s electricity from wind power by 2020. [Online]. Available:http://www.ewea.com/

[2] Commission for Energy Regulation. (2004). Group processing approachfor renewable generator connection applications. [Online]. Available:http://www.cer.ie/

[3] A. Keane and M. O’Malley, “Impact of distribution network constraintson distributed generation capacity,” presented at the 40th Int. Univ. PowerEng. Conf., Cork, Ireland, 2005.

[4] A. R. Wallace and G. Harrison, “Planning for optimal accommodation ofdispersed generation in distribution networks,” presented at the CIRED17th Int. Conf. Electr. Distrib., Barcelona, Spain, May 2003.

[5] A. Keane and M. O’Malley, “Optimal allocation of embedded genera-tion on distribution networks,” IEEE Trans. Power Syst., vol. 20, no. 3,pp. 1640–1646, Aug. 2005.

[6] P. N. Vovos, G. Harrison, A. Wallace, and J. Bialek, “Optimal power flowas a tool for fault level-constrained network capacity analysis,” IEEETrans. Power Syst., vol. 20, no. 2, pp. 734–741, May 2005.

[7] A. Keane and M. O’Malley, “Optimal distributed generation plant mixwith novel loss adjustment factors,” presented at the IEEE Power Eng.Soc. Gen. Meeting, Montreal, Canada, Jun. 18–22, 2006.


[8] P. Chiradeja and R. Ramakumar, “An approach to quantify the technicalbenefits of distributed generation,” IEEE Trans. Energy Convers., vol. 19,no. 4, pp. 764–773, Dec. 2004.

[9] “Total renewable energy resource in Ireland,” ESBI, ETSU, Dublin,Ireland, Tech. Rep. 1997.

[10] ESB National Grid system information. (2005). [Online]. Available:http://www.eirgrid.com/

[11] ESB Networks infrastructure. (2005). [Online]. Available: http://www.esb.ie/esbnetworks/

[12] C. Hird, H. Leite, N. Jenkins, and H. Li, “Network voltage controller fordistributed generation,” Inst. Electr. Eng. Proc. Generation Transmiss.Distrib., vol. 151, no. 2, pp. 150–156, Mar. 2004.

[13] J. Hill, V. Thornley, C. Barbier, A. Maloyd, and A. Oliver, “Innovativevoltage control techniques for optimum connection of renewable and in-termittent generation to distribution networks,” presented at the 40th Int.Univ. Power Eng. Conf., Cork, Ireland, 2005.

[14] KEMA “The contribution to distribution network fault levels from the con-nection of distributed generation,” Dept. Trade Ind., KEMA Ltd., Norway,U.K., Tech. Rep. DG/CG/00027/00/00, May 2005.

[15] All-Island Modelling Project. (2006). [Online]. Available: http://www.allislandproject.org

[16] E. Denny and M. O’Malley, “Wind generation, power system operation,and emissions reduction,” IEEE Trans. Power Syst., vol. 21, no. 1, pp. 341–347, Feb. 2006.

[17] R. Doherty, G. Lalor, and M. O’Malley, “Frequency control in competitiveelectricity market dispatch,” IEEE Trans. Power Syst., vol. 20, no. 3,pp. 1588–1596, Aug. 2005.

[18] Directive 2001/77/EC of the European Parliament and of the Coun-cil. (2001). The promotion of electricity produced from renewable en-ergy sources in the internal electricity market. [Online]. Available:http://www.europa.eu.int

[19] Transmission System Operator Ireland. (2005). Generation adequacy re-port: 2006–2012. [Online]. Available: http://www.eirgrid.com

[20] European Wind Energy Association. (2006). Wind power economics.[Online]. Available: http://www.ewea.com

[21] L. Soder, “Reserve margin planning in a wind-hydro-thermal power sys-tem,” IEEE Trans. Power Syst., vol. 8, no. 2, pp. 564–571, May 1993.

[22] R. Doherty and M. O’Malley, “New approach to quantify reserve demandin systems with significant installed wind capacity,” IEEE Trans. PowerSyst., vol. 20, no. 2, pp. 587–595, May 2005.

[23] S. Lefton, P. Besuner, and G. Grimsrud, “Understand what it really coststo cycle fossil-fired units,” Power, vol. 141, no. 2, pp. 41–42, 1997.

[24] R. Doherty, H. Outhred, and M. O’Malley, “Establishing the role thatwind may have in future generation portfolios,” IEEE Trans. Power Syst.,vol. 21, no. 3, pp. 1415–122, Aug. 2006.

[25] Commission for Energy Regulation Ireland. (2005). Best new entrantprice 2006–CER/05/110. [Online]. Available: http://www.cer.ie

[26] E. Denny, G. Bryans, J. FitzGerald, and M. O’Malley, “A quantitativeanalysis of the net benefits of grid integrated wind,” presented at the IEEEPower Eng. Soc. Gen. Meeting, Montreal, Canada, Jun. 18–22, 2006.

[27] R. Doherty. (May 2006). High level assessment of suitable generation port-folios for the all-island system in 2020. All Island Grid Study Work-Stream2A. Tech. Rep. [Online]. Available: http://www.dcmnr.gov.ie/Energy/

[28] British Wind Energy Association. (2005). [Online]. Available: http://www.bwea.com/energy/

[29] ESB National Grid (Eirgrid). (2005). Statement of charges and paymentsfor ancillary services providers 2005. [Online]. Available: http://www.eirgrid.ie

[30] U.S. Environmental Protection Agency. (2006). Acid rain program.[Online]. Available: http://www.epa.gov

Andrew Keane (S’04) received the B.E. degree inelectrical engineering from the University CollegeDublin, Dublin, Ireland, in 2003. He is currentlyworking toward the Ph.D. degree in the ElectricityResearch Centre, University College Dublin.

His research interests include power systemsplanning, distributed generation, and distributionnetworks.

Eleanor Denny (S’04) received the B.A. degree ineconomics and mathematics and the M.B.S. degreein quantitative finance from the University CollegeDublin, Dublin, Ireland. She is currently working to-ward the Ph.D. degree in the Electricity ResearchCentre, University College Dublin.

Her research interests include the economics ofwind generation.

Mark O’Malley (S’86–M’87–SM’96) received theB.E. and Ph.D. degrees from the University CollegeDublin, Dublin, Ireland, in 1983 and 1987, respec-tively.

He is currently a Professor of electrical engineer-ing at the University College Dublin, and the Directorof the Electricity Research Centre, University CollegeDublin. His current research interests include powersystems, control theory, and biomedical engineering.

Documents

Quantifying the Impact of Connection Policy on Distributed Generation