Effects of PEV Traffic Flows on the Operation of Parking Lots and Charging Stations

Transcription

1 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid Effecs of PEV Traffic Flows on he Operaion of Parking Los and Charging Saions Nilufar Neyesani Member IEEE Maziar Yazdani Damavandi Member IEEE Gianfranco Chicco Senior Member IEEE and João P. S. Caalão Senior Member IEEE Absrac The inroducion of plug-in elecric vehicles (PEVs) in he elecrical sysem is bringing various challenges. The main issue is incorporaing he PEV owner s preferences in he models. One of he main aribues represening he preference of he owners is heir ravel purposes impacing on he raffic flow paern. The PEVs raffic paern defines he required charging schedule of he PEVs and consequenly characerizes he operaion of he charging faciliies such as PEV parking los (s). The deploymen of resources such as PEV requires a deailed modeling of he facors affecing heir operaion. In his regard his paper aims o model he power flow of he PEVs based on heir raffic flow. Differen ravel ypes and purposes are considered for he PEVs raffic modeling. Two ypes of charging infrasrucure (i.e. s and individual charging saions) are considered. The sudy is performed on a disribuion nework caegorized based on he consumpion paerns of he zones. Index erms Charging saion (CS) parking lo () plug-in elecric vehicle (PEV) raffic paern driving behavior. NOTATION Capial leers denoe parameers and small ones denoe variables. Minimum values are underlined and maximum values are overlined. Subscrips Traffic zone Nework nodes Power line Ω Scenario and scenario se h Index of he ime inerval Superscrips Aggregaor Arrived PEVs h/h Baery charging/discharging Coningency Mode Charging Saion Demand Delegaed energy (probabiliy of reserve call) This work was suppored by FEDER funds hrough COMPETE 22 and by Poruguese funds hrough FCT under Proecs SAICT-PAC/4/25 - POCI-- 45-FEDER-6434 POCI--45-FEDER-696 UID/EEA/54/23 UID/CEC/52/23 and UID/EMS/5/23. Also he research leading o hese resuls has received funding from he EU Sevenh Framework Programme FP7/27-23 under gran agreemen no N. Neyesani is wih INESC TEC Poro Porugal (ni.neyesani@gmail.com). M. Y. Damavandi is wih C-MAST Universiy of Beira Inerior Covilhã 62- Porugal (maziar.yazdani.d@gmail.com). J. P. S. Caalão is wih INESC TEC and he Faculy of Engineering of he Beira Inerior Covilhã 62- Porugal and also wih INESC-ID Insiuo Superior Técnico Universiy of Lisbon Lisbon 49- Porugal ( caalao@ubi.p). G. Chicco is wih Poliecnico di Torino Diparimeno Energia Galileo Ferraris corso Duca degli Abruzzi Torino Ialy ( gianfranco.chicco@polio.i). Depared PEVs Energy Exernal Elecric vehicle Energy consumed as he fuel for he vehicle 2 Grid o Vehicle / Inpu/oupu o/from zone or Loss of SOC Power line Parking Lo Reserve Scenario Tariff paid by PEV owners arriving a he TOU Time of use energy price raffic 2 Vehicle o Grid Vacan charging poins in Traffic zone Operaors Δ Change in variable amoun Expeced value of a variable Variables and Parameers Capaciy (kwh) Cos of equipmen depreciaion ( /kwh) Forced ouage rae (%) L Line curren (A) Travel disance beween zones (km) Number Number of PEV saions Acive power (kw) Reacive power (kvar) Reserve (kw) Resisance and reacance of a line (Ω) PEV speed (km/h) u Sae of Charge (kwh) Binary variable Volage (V) α Average ravel ime among zones (h) β SOC increase raio in raffic Γ Charge/discharge rae in saions of / (kw) Δ Φ Time sep (h) Requiremen of PEV owner for minimum SOC SOC loss due o ravel wihin he zone (%) Charge and discharge efficiency (%) Scenario probabiliy (%) Price ( /kwh) I. INTRODUCTION A. Moivaion and Background HE RECENT rends in infrasrucure sudies show he Tendency owards increasing he adopion of plug-in elecric vehicles (PEVs) in he everyday life. The vehicle manufacurers have spen ime and budge on developing various models of PEVs o moivae he end-users o deploy (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

2 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid 2 PEVs insead of radiional vehicles. As he elecric sysem operaors ry o suppor he PEV owners hrough possible incenives he urban sysem planners also need o face wih he preferences of he PEV owners mobiliy in he urban area. Two main soluions for providing he required charging faciliies for he PEVs in he sysem do exis: individual charging saion (CS) and PEV parking lo (). The provision of hese faciliies in he sysem requires a collaboraive planning from elecric nework operaor PEV manufacurers and urban planners. However he widespread aspecs of he problem necessiae a comprehensive sudy hrough each poin of view. Various alernaives for PEV charging are available such as baery swiching. However wih recen progress in elecric vehicle and is baery manufacuring he direc charging of PEVs baeries in he plugged in posiion gained he highes ineres in he field. Consequenly he funcion of PEV s has changed from a poin of baery subsiuion o a medium for he direc ineracion of PEVs wih he grid. On hese bases he PEV owners preferences on how o use heir vehicles would be a criical issue affecing he s operaion. The infrasrucure for he PEVs can be inadequae if he need of he owners is negleced. The PEVs raffic paern is a way of indicaing he PEVs behavior in he sysem. This issue becomes more dominan when he sysem ries o ake advanage of he PEVs as well as providing hem he required charging []. The modeling of he PEVs raffic paern and is incorporaion in he sysem operaion model helps considering he uncerainy of PEVs charging need and he poenial hey bring o enhancing he energy managemen flexibiliy. In his regard his paper proposes a mahemaical model for modeling he PEVs raffic flow. Differen levels of PEVs raffic including he PEVs CS raveling in an area and ravel purposes has been modeled and inegraed wih he sysem operaion model. B. Lieraure Review Alhough he elecric vehicles have been he subecs of many previous sudies here are few sudies ha have focused on he raffic flow of PEVs in a sysem from boh elecric sysem and urban planning poins of view. In [2] various aspecs of elecric mobiliy including power sysem ranspor sysem and he echnology of vehicles for efficien conrol of PEVs in he sysem were sudied. The rips ravelled by PEVs affec heir required energy. The managemen of PEVs power requiremen for hybrid PEVs based on he rips hey ravel was sudied in [3]. On he oher hand when he PEVs are inerconneced o he grid hey will add o he oal load of he nework as hey need elecriciy for heir charging. In [4] he energy needed for he PEVs was considered as a load and was modeled based on he daily disances ha he PEV users ravel. Wih differen raffic behavior he charging in each CS will be differen hus affecing he CS locaion in he grid. In [5] hese effecs were sudied in a planning ime horizon. The auhors in [6] also provided he planning scheme considering he urban raffic flow of he EVs. In addiion [7] derived he behavior of he o be used in is allocaion problem. Moreover in [8] he raffic crieria were added o disribuion sysem planning o provide a coupled elecric and raffic nework plan. In [9] he locaional energy requiremen of PEVs was sudied by considering heir random driving paern. The inerrelaion of elecriciy grid and ransporaion nework for he PEVs case is an imporan issue. In [] a model for he PEVs flee was proposed o be used in he naional energy and ransporaion planning. Also from he urban planning poin of view he allocaion of charging infrasrucures considering he raffic ways and congesions was sudied in []. All of he abovemenioned sudies excep for [5] only considered he grid-o-vehicle (G2V) operaion of PEVs. However he vehicle-o-grid (V2G) poenial of he PEVs also proved o be an effecive componen in he fuure sysem. In [2] in he planning procedure of he disribued energy resources he V2G poenial and raffic paern of PEVs were sudied. Alhough many sudies used he PEVs as heir main concern of sudy only a few conribuions merged he simulaneous concern of PEVs raffic paern effec on he behavior of power sysem componens. C. Conribuions This paper considers boh G2V and V2G operaion modes of he PEVs. Furhermore i proposes a model for represening he operaion of boh s and CSs in he sysem due o differen raffic behaviors. The model considers he PEV owners preferences by aking ino accoun he ravel purposes and zonal raffic paerns as well as modeling he desirable say duraion by considering differen ravel ypes. The requiremens of PEV owners on paricipaing in V2G mode are also considered. The presen model is a comprehensive model ha akes ino accoun all he aspecs of PEV deploymen in he sysem considering ha all s and all CSs are managed by a unique aggregaor. The paper s main conribuions are: To model he power flow due o he raffic flow from and o differen areas (urban and exernal) and o he and CS operaion. To propose a model o derive he s operaion wih differen raffic flow paerns. To model he marke paricipaion of an aggregaor managing s and CSs in energy and reserve markes. D. Paper Organizaion The res of he paper is organized as follows. Secion II explains he main assumpions of he problem. Secion III presens he raffic flow modeling. Secion IV describes he mahemaical opimizaion model for maximizing he aggregaor s profi. Secion V shows he numerical resuls. Secion VI summarizes he main achievemens of he sudy. II. TRAFFIC PATTERN MODEL Wih he PEV soluion fosering he deploymen of he elecric vehicles i becomes criical o examine various aspecs regarding he operaion of s; no only he nework effecs of he bu also heir possible role in fuure elecriciy markes. The poenial of s in paricipaing in he elecriciy markes is considerably dependen on he availabiliy of PEVs in he and heir level of sae of charge (SOC). The aggregaion of he saes of charge of all PEVs in a or in a zone forms he SOC (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

3 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid of he or zone. The SOC is ofen expressed as a percenage of he oal capaciy; in his paper i is expressed in [kwh] o compare i direcly wih he capaciy limis. Only he minimum SOC and maximum SOC are expressed in percenage because hey are applied o variable capaciy values bu he percenage limis remain consan. Higher numbers of PEVs demanding he charge affec he profi of he owners hrough selling energy o he PEV owners. In addiion he willingness of PEVs o paricipae in he V2G mode operaion offering heir level of SOC in he marke increases he profi of s hrough selling energy or reserve o he elecriciy marke. In his regard a proper esimaion of s hourly poenials may help he operaor o design a beer marke paricipaion sraegy. There are various aspecs affecing he poenial of in he marke such as he number of PEVs in he he PEVs say duraion and sae of charge he number of PEVs available for V2G ec. The main facors affecing all hese aspecs are he raffic behavior of he PEVs as well as he preferences of he PEV owners on how hey are going o use heir baery saus. Therefore he role of PEV s raffic paern in PEV s marke sraegy becomes more dominan. As he peneraion of PEVs in he sysem grows he PEV owners will have various opions o charge heir vehicles. As well as he increasing number of individual CSs several s will also be promoed in he sysem. The user s choice beween s and CSs affecs he SOC availabiliy in he. Considering he above consideraions his paper invesigaes he effec of raffic paern of PEVs on he operaion in differen areas of he nework. The area under sudy is divided based on he consumpion paern of he loads on each area which corresponds o: residenial commercial indusrial and complex load (i.e. he combinaion of all load ypes). These consumpion paerns define he ravel purposes of he vehicles driving from one zone o anoher one. In his regard he ravel purposes of vehicles in his sudy are caegorized ino wo main ypes formulaed afer considering he oucomes of he sudy of ravels in he UK repored in [4]:. Travel Type : Residenial o commercial area ravel which includes he home-o-office ravels shopping adminisraive ravels ec. This ype is characerized wih high commues during raffic rush hours and a percenage of longer say duraions for he office ravels. 2. Travel Type 2: Residenial o indusrial area ravel which indicaes he raffic paern based on he indusrial working hour shifs. The ravels of his ype are considered o have low commue and are exended according o 24h working facories. In each area a general-purpose ravel has also been considered named as raffic. This ravel also includes he energy loss of driving on he roads o reach he desinaion. The ravel ypes beween differen areas are shown in Fig.. As i is shown zone #2 and #3 are assigned o indusrial and commercial consumpion paerns respecively and each one only has one ype of ravel as heir arrival/deparure paern. However zone # (residenial) and zone #4 (complex wih a combinaion of all consumpion ypes i.e. residenial commercial and indusrial) have boh ypes of ravel. Moreover in each zone i is considered ha some of he PEVs have a desinaion oher han he specified zones and ravel o anoher area which is called Exernal area in his sudy. Some vehicles also ener from he Exernal area o he defined zones. This assumpion is necessary because i is no realisic o consider ha PEVs are bound o move wihin a specified area. Table I represens he varians ha may be found in he sudies on elecric vehicles concerning charging ype operaion mode and ways of managing s and CSs. The managing eniy may be a CS or /CS aggregaor or an individual eniy such as he PEV user (for a Home CS) or an exernal operaor (for Home or CSs). Under his general scheme i is possible o idenify combined soluions in which for example independen operaors manage he individual s while all CSs could be managed by an aggregaor or oher soluions. The indicaions provided in Table I refer o he characerisics of he sysem under analysis in which he s and he CSs are managed in an aggregae way by he same eniy and here is no Home CS. I is assumed ha each zone is equipped wih various numbers of s and CSs in he area. Each provides G2V while V2G is available in an opional way i.e. subec o an addiional fee. Each CS provides only G2V faciliy; hence i is reaed as a load in he modeling. TABLE I. PRESENCE AND MANAGEMENT OF S AND CSS (THE INDICATIONS REFER TO THE OPTIONS USED IN THIS PAPER; THE GRAYED CELLS ARE NOT RELEVANT) aggregaed individual charging managemen managemen mode ype unique (for s only CSs only PEV user operaor s+css) G2V yes V2G opional G2V Home CS V2G G2V yes CS V2G Environmen #3: Commercial CS CS #: Residenial #2: Indusrial CS CS #4: Complex combinaion of residenial commercial and indusrial usage Travel Type (T ) Travel Type 2 (T 2 ) Combinaion of boh ravel ypes (T &T 2) Fig.. Traffic flow of differen ravel ypes beween consumpion areas (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

4 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid 4 In each area and based on he ravel ype he PEV owners preferences on using he or CS as well as heir charging requiremens are also considered in he model. The PEVs can choose beween and CS based on heir preferred say duraion. Moreover hey can deermine heir minimum deparure SOC requiremen. For hose vehicles ha choose he choice beween G2V and V2G mode is also considered. This means ha no all he PEVs in he are obliged o paricipae in V2G mode bu if hey are willing o ake par hey will receive an exra paymen for heir baery degradaion. III. TRAFFIC FLOW MODELING The raffic flow analysis is carried ou by considering successive ime inervals of given duraion. The variable is used o represen a ime inerval while he variable is used o represen a scenario of arrival and deparure o each zone and CS. I is assumed ha he arriving/deparing number of PEVs and capaciy o/from each zone are based on scenarios and are considered as parameers. The raffic flow o/from each zone is depiced in Fig. 2. However afer enering he zone he pariioning of PEVs beween and CS would be differen in each zone and is based on he PEVs preferences in each zone; hence he numbers of PEVs are considered as variables (e.g. by considering he scenario he variable represens he number of PEVs arriving a ime inerval in he s locaed in zone i). When he paern of arrival/deparures in each zone is deermined he flow of he energy ha goes wih he vehicles should also be modeled. In his paper i is assumed ha in each zone he vehicles have he choice beween and CS based on heir preferences for geing charged or willingness of geing discharged. I is also assumed ha he PEVs ha will ener in he need charging oherwise hey would no be willing o pay he parking fee. Moreover he urban driving also causes energy loss in PEV baeries. Therefore he model also considers he energy loss due o driving in each zone. The energy flows of he PEVs as well as he ineracion wih he grid are shown in Fig. 3. In he figure he oal SOC of arriving vehicles from oher zones and he PEVs arriving o he zone from he Exernal area is considered as he arrival SOC of he zone ( ). The arrived SOC o he zone hen will be spli beween he and he raffic. This means ha when he vehicles arrive o he zone hey go direcly o he or have oher desinaions and ravel purposes in he zone. While being in he area each vehicle has he opion of geing charged hrough fascharging faciliies provided by individual CSs in he zone. Thus anoher inpu is added o he module which is he average power corresponding o he required energy inpu for charging he vehicles in he CS. On he oher hand he PEVs ha park in he provide V2G mode as well as G2V. This assumpion will resul in inpu and oupu energy requiremen of ( ). The inpu/oupu power of he and inpu power o he CS will change he arrival SOC; hence he deparing SOC from he zone is compued from he deparure SOC of and. N in i in Ex N i ni ni n i ni ni ni Fig. 2. Represenaion of he raffic flows of moving PEVs. soc in i soc i in Ex soc i soc soci i soc soc i in pi loss pi in ou pi pi dep dep i soc i Fig. 3. Represenaion of he charging flows of moving PEVs. N ou i ou Ex N i soc soc i ou i ou Ex The raffic behavior of he PEVs in each area (i.e. ravel ype) deermines he arrival SOC o he zone and he PEV owners preferences and ravel purposes esablish he arrival SOC o he and CS. In addiion he marke sraegy of he affecs is operaion and consequenly affecs he s deparure SOC. The deparure SOC of one zone is he inpu SOC of anoher zone inerfering wih he possible marke sraegy of oher s in oher zones. This paper provides he model o calculae he power flow based on he raffic flow and grid ineracion of he elecric vehicles. The proposed model of he PEV s power flow based on heir raffic flow is mahemaically presened below. A. raffic model Referring o Fig. 3 in order o consruc he raffic model in each zone i is necessary o calculae he ineracions of he PEVs beween zones ha are deermined by heir number capaciy and SOC. In his subsecion he required formulaion for hese calculaions is presened. The duraion of he ravel is deermined by he vehicle s speed and he disance ha is raveled (). Hence a PEV depared from zone i a ime inerval will arrive o zone a h wih a loss of fuel based on he ravel s speed and disance as in (2). α i = Li S () i in ou ou Fuel h = + α i soc i h = soc i N i Li Pi (2) The arrived PEVs o each zone deermine he oal capaciy of he zone. The compued hourly arrival/deparure SOC o/from each zone should no exceed he oal available capaciy in ha zone (which is deermined by scenarios) as in (3) and (4). The inenion of hese wo consrains is o limi he compuaion of SOC ransfer beween zones wih he raffic paern scenarios. in in soc i C i (3) ou ou soci Ci (4) On enering he zone he PEV owners have several opions: o go direcly o he o go direcly o he CS and ravel in he zone hen go o or CS. I is assumed ha he PEVs firs choice in indusrial residenial and complex areas (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

5 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid 5 are o go o he as he raffic paern in hose areas suggess longer say duraions and hey can benefi from paricipaing in he V2G mode. On he oher hand based on he high commue and shorer say duraion paern in he commercial area he firs choice of he PEVs in he commercial area is o go o he CS as heir main requiremen is fas charging. Anyhow regardless of he PEVs choice inside he zone he oal number of vehicles arriving o or should be equal o he oal number of PEVs enering each zone from oher zones and from he Exernal area (5). The same reasoning is used for he arrival capaciy of he PEVs in and. in Ex in n + n = N + N (5) i i i i Alhough he PEVs may prefer o go o he raher han CS he limied number of charging poins in he may no correspond o all he vehicles ha need a parking space for charging. Thus in each ime inerval he available number of vac (vacan) charging poins in he in each zone ( ni ) is compued from (6). Then he arrival number of PEVs o he is derived from (7) based on he availabiliy of he vacan charging poins in he. vac n = NS n (6) i i i i = vac in Ex in vac ni if Ni N i ni in Ex in in Ex in vac Ni + N i if Ni + N i ni n + > (7) In (7) i is shown ha if he oal number of vehicles arriving o a zone in each ime inerval is less han he vacan charging poins in he hen he oal arrived PEVs o he zone can ener he. Oherwise only he vacan charging poins will be filled wih he newly arrived PEVs. Alhough he number of PEVs in he is deermined by (7) he calculaion of he arrival SOC is differen due o he charging/discharging of he in preceding ime inervals and in differen zones. In his regard he arrived SOC is compued from soc i = ar in Ex in vac Δ soci k if Ni + N i > ni k (8) in Ex in in Ex in vac SOCi + soc i if Ni + N i ni In his sudy a new approach is formulaed o calculae he variable level of SOC of he arrival vehicles o he. In his approach considering he scenario and he ime inerval he binary variable ui k is defined o represen he presence of a vacan charging poin in node k (binary variable = ) or no (binary variable = ) in he s of zone i. The oal number of vacan charging poins in he and he raio of oal arrival SOC and PEV numbers o he zone are hen deermined as in (9) and (). vac ui k = ni (9) k ar Δ soci k = in Ex in in Ex in () ( SOCi + soc i ) ( Ni + N i ) ui k The arrival paern of PEVs has been modeled wih arrival scenarios and PEVs preference on choosing. However esimaing he deparing PEVs from he zone canno be considered in he same way. This issue is more criical when compuing he deparure SOC. The reason is ha he arrival SOC is also affeced by he charging/discharging in he or CS. In his regard he following approach is adoped o calculae he number capaciy and SOC of he PEVs. Wihou loss of generaliy he formulaion is illusraed only for he SOC. In () i is shown ha he deparure SOC of and will form he deparure SOC ha goes ou from a zone o he Exernal area or oher zones. The equaions (2) and (3) show ha a each ime inerval he depared SOC of he and are proporional o he level of SOC in he or muliplied by he number of ougoing PEVs from he zone and he oal number of exising PEVs in he zone. soc + soc = soc = soc + soc () soc ou Ex ou i i i i i = soc i i. ou Ex ou Ni + Ni N + N N N + N + N in Ex in ou Ex ou i h i h i h i h i = i = h soc = soc i i. ou Ex ou Ni + Ni N + N N N + N + N in Ex in ou Ex ou i h i h i h i h i = i = h (2) (3) Afer calculaing he deparure SOC of he and i should be deermined wha share of he deparure SOC goes o anoher zone and how much ravels o he Exernal area. This allomen is done using (4) and (5). ou Ex ou Ex ou Ex ou soci = Ni ( Ni + Ni ) soc (4) i ou ou ou Ex ou soci = Ni ( Ni + Ni ) soc (5) i B. raffic model Afer enering he zone he PEVs may lose heir SOC due o driving wihin he zone area or hey may increase heir level of SOC wih enering he CS o charge. Therefore he SOC balance wihin he area is based on (6) where he SOC of arrived and depared vehicles are considered as well as he inpu power hrough he CSs and he efficiency aking ino accoun he power loss due o driving wihin he area (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

6 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid 6 in cha soc = i SOC + i soc + i p i η Δ i = > (6) loss + soci soci soci The loss of SOC due o driving in he area is compued wih loss soci = soci λi (7) The maximum and minimum level of SOC in he area are defined by EV EV SOC i ci soc i SOC i ci (8) I should be noed ha he maximum ineced energy o he PEVs hrough he CSs in a ime inerval canno be higher han he remainder of he capaciy of he PEVs in he area. Thus he oal number of CSs he oal number of vehicles in each ime inerval and he available capaciy of he vehicles in he area are compared o deermine he maximum limi for power inecion ino he CSs. The coefficien β i is a facor showing he increase in willingness o charge when lower levels of SOC remain in he baery. p in i (9) min Γ CS i NS CS i Γ CS i n i C { ( i soc i )β i Δ} In each ime inerval he oal number of vehicles in he area follows he balance in (2). The same balance can be used for calculaing he capaciy of he area. n = N + n + n n (2) i i = i > i i C. raffic model The effec of he raffic behavior on s SOC has been addressed in deail in [3] and is also considered in his sudy. In each ime inerval he SOC of he is affeced by he charging and discharging of he PEVs as well as he arrival/deparure SOC of he PEVs (2). in cha soc = i SOC + i soc + i p i η Δ i = > (2) ou dcha pi Δ ηi + soci soci In each ime inerval he oal inpu power of he is consrained by he oal number of vehicles in he muliplied by he charging rae of he (22). The maximum possible oupu power of he is limied by he oal number of PEVs in he muliplied by he discharging rae of he charging poin equipmen and by he deparure SOC requiremens of he PEV owners (23). The PEVs preferences on heir deparure SOC requiremens are deermined by he coefficien φ. i p in i Γ i n i (22) p ou i min Γ i n { i soc i φ i Δ} (23) When he is paricipaing in he reserve marke as well as he energy marke he summaion of s energy and reserve oupu should be less han he minimum of s possible charging capaciy and he available SOC in he o be offered in he marke (24). { } (24) p ou i + r ou i min Γ i n i soc κ i Δ i Like he urban area he s SOC is limied o he maximum and minimum percenage of oal s capaciy (25). EV EV SOC i ci soci SOC i ci (25) The number of vehicles in he in each ime inerval is calculaed from (26) and should be less han he oal number of charging poins in he as in (27). n = N + n + n n (26) i i = i > i i ni NSi (27) IV. AGGREGATOR OBJECTIVE MODEL I is assumed a single aggregaor is responsible for managing he energy flows in he sysem. Thus he obecive funcion of he aggregaor is o maximize he aggregaor s profi hrough marke ineracions and selling energy o he PEVs in s or CSs. As shown in (28) he aggregaor buys and sells energy from/o he energy marke (upsream nework) wih energy price bu rades wih s based on he PEV ariffs. Based on he obecive he aggregaor can make profi hrough selling energy o he upsream nework reserve marke paricipaion s ariff and selling energy o PEVs in he s. I is assumed ha he energy sold o PEVs in he CSs has he same price as he energy marke price while he G2V energy is sold o PEVs in he s a a price of π i lower han he marke price. Selling o he PEV s a lower price is an incenive-based sraegy aken by he owner o encourage he paricipaion of PEVs in he. This approach in he long erm would also help he owner o recover is mainenance coss [7]. The ariff and G2V/V2G prices are deermined by he manager (he aggregaor in his case). Agg Profi = Agg Cosk + ρ ( Profii Cosi ) = k i Aggin Aggou E Δ( pk pk ) π (28) k ou R del E Tariff i i i i i ( ρ ) +Δ ρ r π + r ρ π + n π p + r Cd r p ou del del Con del V2G i i i i i i i i i π in G 2 V ou V 2G + i i i i p π ρ FOR π r ρ π The sysem power flow consrains are added o he obecive funcion. However o assign he energy ineracion of each or CS in each node he following approach is adoped. Firs for each zone he oal or urban behavior is deermined using he models in Secion III. Then he expeced share of each grid node k (wheher or urban) from inpu/oupu power is calculaed proporional o he number of PEVs in he or CS (29)-(3). The same applies for he expeced reserve of he (32). ) (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

7 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid ( ) pˆ = ρ p N N (29) in in oal k i ik i ( ) pˆ = ρ p N N (3) ou ou oal k i ik i ( ) pˆ = ρ p N N (3) in in CS CS oal k i ik i ( ) r = ρ r N N (32) ˆ oal k i ik i The acive and reacive power balance of he sysem is shown in (33) and (34) respecively. The power flow equaion is formed as in (35) and is linearized using he echnique indicaed in [7]. ( ) 2 p p + p p + R i = (33) P p p p = (34) Aggin Aggou Line Line k k k k kk kk kk l l D in ou in ˆ ˆ ˆ k + k k + k r ˆ ( ) oal k ρ ri Nik Ni 2 Line Line Line Line vk 2 Rkk ( pkk pk k ) + Xkk ( qkk qk k ) Zk k ikk vk = (35) V. NUMERICAL RESULTS The IEEE 37 bus radial disribuion es sysem [4] is used o model he area under sudy wih various consumpion paerns as shown in Fig. 4. The boundaries for each zone and he locaion of s and CSs are indicaed in he grid. Previous sudies [5] and [6] have deermined he maximum possible inecion from each node ino he grid. This limi is considered for he deerminaion of he maximum number of charging poins in each. The inpu daa for PEVs raffic paern (i.e. arrival/deparure number capaciy and SOC of PEVs) are considered o be sochasic and modeled wih differen scenarios for each zone. The uncerainy characerizaion of he PEVs is modeled using he approach in [6] wih 3 minues ime frame. The raffic behavior of PEVs in each zone is adaped o he consumpion paern in each zone based on various raffic behavior sudies including [7-9]. The ravel ypes assumed in his sudy are shown in Fig. 5. The ime inerval used for he analysis is 3 minues. However for he sake of beer undersanding he figures are shown in a 24-hour frame. The exchanges beween zones are shown in Fig. 6 Fig. 7 and Fig. 8. Based on Fig. he ravel from zone # o #2 is ravel ype and follows he same paern depiced in Fig. 5. As shown in Fig. 6 he depared capaciy from zone # will reach zone #2 wih a ime delay. However he depared SOC from zone # will reach zone #2 boh wih he ime delay and he loss of SOC level due o consuming he baery charge during he driving. Moreover he driving paerns of PEVs in each zone are in accordance wih he consumpion paern of he zone. I shows ha he PEVs leave he residenial area a he beginning of he working hours and come back a he end of he day. However as i is assumed ha he facories in he indusrial zone work for hree work shifs he arrival/deparure hours are more spread comparing o raffic paern of residenialcommercial ravels. Number of PEVs Arrival Type Deparure Type Arrival Type 2 Deparure Type Fig. 5. Expeced number of PEVs in arrival/deparure of each ravel ype. Fig. 4. IEEE 37 bus nework wih s and CSs divided based on he usage paerns. Capaciy and SOC of PEVs Depared SOC from #2 o # Arrived SOC from #2 o # Depared SOC from # o #2 Arrived SOC from # o #2 Arrived Capaciy from #2 o # Depared Capaciy from #2 o # Arrived Capaciy from # o #2 Depared Capaciy from # o # Fig. 6. Capaciy and SOC flow beween # and # (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

8 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid Capaciy and SOC of PEVs Fig. 7. Capaciy and SOC flow beween # and #3. Capaciy and SOC of PEVs Depared SOC from #3 o # Arrived SOC from #3 o # Depared SOC from # o #3 Arrived SOC from # o #3 Arrived Capaciy from #3 o # Depared Capaciy from #3 o # Arrived Capaciy from # o #3 Depared Capaciy from # o # Depared SOC from #4 o # Arrived SOC from #4 o # Depared SOC from # o #4 Arrived SOC from # o #4 Arrived Capaciy from #4 o # Depared Capaciy from #4 o # Arrived Capaciy from # o #4 Depared Capaciy from # o # Fig. 8. Capaciy and SOC flow beween # and #4. Daa for he day-ahead marke are obained from he Spanish elecriciy marke [2]. I is assumed ha in he only one kind of charging poin is used. Based on [2] and [22] i is assumed ha he s have he medium-charging poin a a charging rae of kw per hour. On he oher hand he CSs are equipped wih fas-charging faciliies so ha he vehicles would no need o say long for geing charged. The proposed model is formulaed as a mixed ineger linear programming (MILP) problem and is solved wih GAMS by using he CEX2 solver wo 3.7-GHz Inel six-core processors wih GB of RAM running 64-bi Windows. The response ime of he program is abou 35 minues which does no affec he applicabiliy of he model. The resuls for energy ineracion of PEVs in each zone are shown in figures 9 o 2. In all he figures he oal energy ineracion of he s and CSs in each zone is shown as well as he energy loss in he urban driving. As i can be seen he energy ineracion of he is in accordance wih he availabiliy of he PEVs in he zone based on each zone s raffic paern. In Fig. 9 in he residenial area he is charging during early hours in he morning when he energy price is lower while during lae hours of he day more charging is done wih CSs. A hour 2 he inecs ino he grid as a compromise of energy and reserve price. The analysis of he energy ineracions in each zone can be performed wih consideraion of he raffic behavior. In zone #2 wih he indusrial paern he raffic has a low commuing paern. Moreover he hree working shifs cause higher say duraions and higher number of PEVs remaining in he area. However charging hrough he s is mainly posponed for he lower price hours (from o 7 a.m.). On he oher hand in zone #3 where he commuing is high boh and CS end o charge he PEVs based on heir requiremens before heir deparure. As higher numbers of PEVs are raveling in zone #3 higher charging hrough CSs happens in his zone. Besides higher level of urban loss is observed in his zone comparing o zone # and #2. Energy (kwh) Fig. 9. Energy ineracions in #. Energy (kwh) Fig.. Energy ineracions in #2. Energy (kwh) Energy Ineracion of Energy Loss in Fig.. Energy ineracions in #3. Energy (kwh) Energy Ineracion of Energy Loss in Fig. 2. Energy ineracions in #4. Energy Inecion of CS Energy Ineracion of Energy Loss in Energy Inecion of CS Energy Ineracion of Energy Loss in Energy Inecion of CS Energy Inecion of CS (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

9 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid For zone #4 (Fig. 2) a combinaion of all ravel ypes is presen in he zone bu he charges he PEVs only in he low-price hours. The reason is ha based on he ravels beween zones he SOC of he PEVs ha reach zone #4 is affeced by he charging ha hey receive in oher zones. Anoher affecing facor is he number of charging poins versus CSs in each zone. In zones # and #2 due o low commuing and longer say duraions he number of charging poins is higher han in zones #3 and #4 where due o high commuing paern he PEVs prefer o use fas charging of CSs. The oal reserve provision of s in each zone is shown in Fig. 3. As can be seen he reserve provision of zone #4 has smooher paern comparing o he oher zones. This assures he profi of s in zone #4 despie of no selling energy o he PEVs. The resuls show ha he s in zones # and #2 have higher amoun of reserve provision due o longer say of PEVs in he parking on he oher hand he availabiliy of he reserve is bound o he specific hours. For zone #3 he s profi is mainly hrough selling energy o s due o shorer say raher han providing reserve o he sysem. Reserve Provision (kwh) Reserve of #4 Reserve of #3 Reserve of #2 Reserve of # Toal Capaciy of Toal SOC of Reserve Price Fig. 3. Reserve provision for he sysem hrough s in all zones. VI. CONCLUSIONS In his sudy various consideraions regarding he energy flown wih he raffic of PEVs in a disribuion sysem have been sudied. Two ypes of PEV infrasrucures i.e. s and CSs have been added o he grid o mee he PEV owners preferences. s and CSs are managed by he same eniy. The resuls invesigaed he marke benefi of he energy provider (aggregaor) operaing hose componens. I was concluded ha as differen raffic paerns give differen behavior o he s in each zone differen levels of flexibiliy will be provided for he aggregaor o exploi in is marke sraegy. The raffic behavior of he residenial and indusrial zones gave he aggregaor he poenial of reserve marke paricipaion while he commercial zone helped he aggregaor o make profi hrough selling energy o high commue PEVs on heir shor say. Alhough he problem invesigaed he aggregaed operaion of he s in differen zones he resuls show he possible cross effecs of each s sraegy on charging he PEVs on he oher s behavior. Moreover i gives he insighs for invesmen decision making for he insallaion of he based on he consumpion paern Reserve Price (Cen/kWh) of he area. I can be concluded ha if an invesor decides o insall a in a residenial area i is beneficial if he reserve marke paricipaion is possible for he operaor. On he oher hand he resuls sugges ha deploying a combinaion of all possible PEV ravels (as in he complex zone) gives a smooh profile o he sysem operaor and simplifies he operaion of he sysem. Finally he oucomes of his sudy show ha he proposed model is efficien and can be used for providing he guidelines on he PEV planning as well as he operaion of a sysem wih higher peneraion of PEVs. The sudies based on his model can be adoped o evaluae he economic aspecs of he along wih he PEV impacs on he nework. Moreover as i also considered he raffic flow of he PEVs beer insighs for urban planning of s insallaion can be achieved. REFERENCES [] N. S. Pearre W. Kempon R. L. Guensler V. V. Elango Elecric vehicles: How much range is required for a day s driving Transporaion Research Par C: Emerging Technologies vol. 9 no. 6 pp [2] M. D. Galus R. A. Waraich F. Noembrini K. Seurs G. Georges K. Boulouchos K. W. Axhausen and G. Anderson Inegraing Power Sysems Transpor Sysems and Vehicle Technology for Elecric Mobiliy Impac Assessmen and Efficien Conrol IEEE Trans. Smar Grid vol. 3 no. 2 pp [3] Q. Gong Y. Li and Z. Peng Trip-Based Opimal Power Managemen of Plug-in Hybrid Elecric Vehicles IEEE Trans. Vehicular Technology vol. 57 no. 6 pp [4] K. Qian C. Zhou M. Allan and Y. Yuan Modeling of Load Demand Due o EV Baery Charging in Disribuion Sysems IEEE Trans. Power Sys. vol. 26 no. 2 pp [5] G. Wang Z. Xu F. Wen and K. P. Wong Traffic-consrained muli obecive planning of elecric-vehicle charging saions IEEE Trans. Power Delivery vol. 28 pp [6] Z. Liu F. Wen and G. Ledwich "Opimal Planning of Elecric-Vehicle Charging Saions in Disribuion Sysems" IEEE Trans. on Power Delivery vol. 28 no. pp [7] N. Neyesani M. Y Damavandi M. Shafie-khah J. Conreras and J.P.S. Caalao Allocaion of Plug-In Vehicles' Parking Los in Disribuion Sysems Considering Nework-Consrained Obecives IEEE Trans. on Power Sysems vol. 3 no. 5 pp [8] W. Yao e al. "A Muli-Obecive Collaboraive Planning Sraegy for Inegraed Power Disribuion and Elecric Vehicle Charging Sysems" IEEE Trans. on Power Sysems vol. 29 no. 4 pp [9] M. E. Khoday L. Wu and M. Shahidepour Hourly Coordinaion of Elecric Vehicle Operaion and Volaile Wind Power Generaion in SCUC IEEE Trans. Smar Grid vol. 3 no. 3 pp [] D. Wu and D. C. Alipranis Modeling ligh-duy plug-in elecric vehicles for naional energy and ransporaion planning Energy Policy vol. 63 pp [] N. Sahaye and S. Kelley An approach for he opimal planning of elecric vehicle infrasrucure for highway corridors Transporaion Research Par E no. 59 pp [2] G. Cardoso M. Sadler M.C. Bozchalui R. Sharma C. Marnay A. Barbosa-Póvoa P. Ferrão Opimal invesmen and scheduling of disribued energy resources wih uncerainy in elecric vehicle driving schedules Energy vol. 64 pp [3] N. Neyesani; M. Yazdani Damavandi; M. Shafie-khah; A. Bakirzis; J. P. S. Caalao "Plug-in Elecric Vehicles Parking Lo Equilibria wih Energy and Reserve Markes" IEEE Trans. on Power Sysem doi:.9/tpwrs [4] W. H. Kersing Radial disribuion es feeders Proc. IEEE Power Eng. Soc. Winer Meeing 2 vol. 2 pp [5] Z. Liu F. Wen and G. Ledwich Opimal Siing and Sizing of Disribued Generaors in Disribuion Sysems Considering Uncerainies IEEE Trans. Power Del. vol. 26 no. 4 pp (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.

10 This aricle has been acceped for publicaion in a fuure issue of his ournal bu has no been fully edied. Conen may change prior o final publicaion. Ciaion informaion: DOI.9/TSG IEEE Transacions on Smar Grid [6] Y. G. Hegazy M.M. Ohman W. El-Khaam and A. Y. Abdelaziz Opimal sizing and siing of disribued generaors using Big Bang Big Crunch mehod Proc. Inernaional Universiies Power Engineering Conference (UPEC) pp. -6 Sep. 24. [7] G. Pasaoglu D. Fiorello L. Zani A. Marino A. Zubaryeva and C. Thiel Proecions for Elecric Vehicle Load Profiles in Europe Based on Travel Survey Daa European Comm. Rep. EUR EN 23 [Online]. Available: hps://seis.ec.europa.eu/publicaions/rc-seisrepors. [8] G. Pasaoglu D. Fiorello A. Marino G. Scarcella A. Alemanno A. Zubaryeva and C. Thiel Driving and parking paerns of European car drivers - a mobiliy survey European Comm. Rep. EUR EN 22 [Online]. Available: hps://seis.ec.europa.eu/publicaions/rcseis-repors. [9] R. van Haaren Assessmen of elecric cars range requiremens and usage paerns based on driving behavior recorded in he Naional Household Travel Survey of 29. Sudy of he Solar Journey USA. Earh and Environmenal Engineering Deparmen Columbia Universiy Fu Foundaion School of Engineering and Applied Science New York December 2. [2] Red Elécrica de España Maximum average hourly power demand and daily energy [Online]. Available: hp:// [2] A. Briones J. Francfor P. Heimann M. Schey S. Schey J. Sm Vehicle-o-grid(V2G) power flow regulaions and building codes review by he AVTA Idaho Naional Lab. Idaho Falls ID Sep. 22 [Online]. Available: hp:// [22] Y.K. Jemal Plug-in elecric vehicle charging impacs on power sysems M.Sc. Thesis Dep. of Energy and Environmen Chalmers Univ. of Tech. Goeborg Sweden 2. Nilufar Neyesani (S 9 M 4) received he M.Sc. degrees in elecrical engineering from Iran Universiy of Science and Technology Tehran Iran in 2 and Ph.D. from Universiy of Beira Inerior (UBI) Covilha Porugal in 26. She is currenly a posdocoral fellow a INESC TEC Poro Porugal. Her research ineress include elecric vehicles smar grids energy and reserve markes power sysem opimizaion mulienergy sysems and energy hub. Maziar Yazdani Damavandi (S 8-M 4) received he M.Sc. and PhD degrees in elecrical engineering from Tarbia Modares Universiy Tehran Iran in 2 and 25 respecively. He is currenly a pos-docoral fellow of COMPETE 22-Porugal proec ESGRIDS (Enhancing Smar GRIDs for Susainabiliy) a C-MAST/UBI Covilha Porugal. His research ineress include opimizaion models in power sysem sudies and muli-energy sysem modeling. Gianfranco Chicco (M 98 SM 8) received he Ph.D. degree in elecroechnics engineering from Poliecnico di Torino (PdT) Torino Ialy in 992. He is currenly a Professor of elecrical energy sysems wih he Energy Deparmen PdT. His research ineress include power sysem and disribuion sysem analysis energy efficiency muligeneraion load managemen arificial inelligence applicaions and power qualiy. Prof. Chicco is an Edior of he IEEE Transacions on Smar Grid and a Member of he Ialian Associaion of Elecrical Elecronic and Telecommunicaions Engineers. João P. S. Caalão (M 4-SM 2) received he M.Sc. degree from he Insiuo Superior Técnico (IST) Lisbon Porugal in 23 and he Ph.D. degree and Habiliaion for Full Professor ("Agregação") from he Universiy of Beira Inerior (UBI) Covilha Porugal in 27 and 23 respecively. Currenly he is a Professor a he Faculy of Engineering of he Universiy of Poro (FEUP) Poro Porugal and Researcher a INESC TEC INESC-ID/IST-UL and C- MAST/UBI. He was he Primary Coordinaor of he EU-funded FP7 proec SiNGULAR ("Smar and Susainable Insular Elecriciy Grids Under Large- Scale Renewable Inegraion") a 5.2-million-euro proec involving indusry parners. He has auhored or coauhored more han 55 publicaions including 85 ournal papers 325 conference proceedings papers 3 book chapers and 4 echnical repors wih an h-index of 32 and over 435 ciaions (according o Google Scholar) having supervised more han 5 posdocs Ph.D. and M.Sc. sudens. He is he Edior of he books eniled Elecric Power Sysems: Advanced Forecasing Techniques and Opimal Generaion Scheduling and Smar and Susainable Power Sysems: Operaions Planning and Economics of Insular Elecriciy Grids (Boca Raon FL USA: CRC Press 22 and 25 respecively). His research ineress include power sysem operaions and planning hydro and hermal scheduling wind and price forecasing disribued renewable generaion demand response and smar grids. Prof. Caalão is an Edior of he IEEE TRANSACTIONS ON SMART GRID an Edior of he IEEE TRANSACTIONS ON SUSTAINABLE ENERGY an Edior of he IEEE TRANSACTIONS ON POWER SYSTEMS and an Associae Edior of he IET Renewable Power Generaion. He was he Gues Edior-in-Chief for he Special Secion on "Real-Time Demand Response" of he IEEE TRANSACTIONS ON SMART GRID published in December 22 and he Gues Edior-in-Chief for he Special Secion on "Reserve and Flexibiliy for Handling Variabiliy and Uncerainy of Renewable Generaion" of he IEEE TRANSACTIONS ON SUSTAINABLE ENERGY published in April 26. Since May 27 he is he Corresponding Gues Edior for he Special Secion on "Indusrial and Commercial Demand Response" of he IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS. He was he recipien of he 2 Scienific Meri Award UBI-FE/Sanander Universiies and he 22 Scienific Award UTL/Sanander Toa. Also he has won 4 Bes Paper Awards a IEEE Conferences (c) 27 IEEE. Personal use is permied bu republicaion/redisribuion requires IEEE permission. See hp:// for more informaion.