1
Current Energy Management Technologies Research in China Considering EVs Integration Qinglai Guo, Member, IEEE, Hongbin Sun*, Member, IEEE, Yao Wang, Zhengshuo Li, Boming Zhang, Fellow, IEEE
Abstract — Some latest research works on EVs integration in China are presented in this paper. A future energy management framework supporting large-scale EVs fleets is proposed, involving five related topics: 1) modeling of EV swarms charging load based on naturalistic travel patterns; 2) global power flow based assessment and early warning technologies considering real-time traffic information; 3) emergency control by regulating EVs charging infrastructures; 4) optimized coordination with EVs and renewable energies; 5) interaction between power system and transportation system. Finally, some common features of the above researches are surveyed. Index Terms —Electric Vehicle, Intelligent Transportation Systems
T
Energy Management,
I. INTRODUCTION
he development of Electric Vehicles (EVs) and charging infrastructures is rapid in recent years. The Pike Research predicts that 5.2 million EVs will have been sold worldwide by the end of 2017 [1]. Many projects about EVs are broadly under developing all over the world, such as the “Partnership for a New Generation of Vehicles” in US and “National Electric Vehicle Development Program” in German. In China, the new energy vehicles are also determined to be one of the seven strategic emerging industries, and 87 charging stations (or battery-swapping stations) as well as 7031 charging poles have been built at present. The energy of EVs comes from the power system and largescale EV fleets may be the great challenge for the future power grid. For the power system, the high penetration of EVs may potentially cause great negative influence to operation security, such as load peak or voltage sag. On the other hand, if we can’t handle the coordination problem between the grid and the EVs, the power grid operating constraints may become the new bottleneck restricting the development of EVs. Therefore, V2G (Vehicle to Grid) studies have been spotlighted as a common interest for both power system field and transportation field.
This work was supported in part by MOST (Ministry of Science and Technology) of China (No.2010DFA72760, No. 2011AA11290, No. 2012AA050211), National Science Foundation of China (50807025), National Science Fund for Distinguished Young Scholars (51025725) and Tsinghua University Initiative Scientific Research Program Hongbin Sun* is the corresponding author (e-mail:
[email protected]). Qinglai Guo, Hongbin Sun, Yao Wang, Zhengshuo Li and Boming Zhang are all with the Department of Electrical Engineering, State Key Laboratory of Power Systems, Tsinghua University, Beijing 100084, China.
978-1-4673-2729-9/12/$31.00 ©2012 IEEE
The basic concept of V2G is that EVs will feed the electrical energy back to the power grid when parking [2]. However, in a broader viewpoint, V2G involves the bi-directional energy flow [3]. The implementation of V2G will support the integration of EVs. For instance, V2G offers opportunities to optimize the utilization of the renewable generation [4, 5]. The fluctuation in the output of the renewable generators can be smoothed by coordinated controlling the charging of EVs. In addition, ancillary services can be provided by the implementation of V2G [6]. As a member of CERC-CVC (US-China Clean Energy Research Center – Clean Vehicle Collaboration), our group has taken some V2G related research tasks since 2010. Some works have been reported in details before [4, 7] and more works are ongoing now. In this paper, we will organize and briefly introduce these works from a different viewpoint, which is how to integrate large-scale EVs into the energy management framework. II. FRAMEWORK The power system is an ultra-large system where energy flows, which must satisfy the balance of power in all the times and all the spots. On the other hand, EV is a kind of storage which energy diminishes when moving. As long as the grid and EVs are in connection, two-way energy flow is available. The energy exchange depends not only on the location of connecting, but also the operation state of the power grid. It is a great challenge for the future smart grid to modeling, monitoring, optimizing and controlling EVs together with other elements, which is actually an energy management problem. An energy management system (EMS) is a computer-aided tool used by operators of electric utility grids to monitor, control, and optimize the performance of the power system. EMS is so essential for power system operation that it is usually regarded as the brain. Electric vehicles connected into the power grid are a kind of new “source” which features are totally different from the conventional generations and demands. So it is important to integrate this new member into the existed EMS framework, and no doubt, many works from modeling to controlling are supposed to been carried out, as shown in Fig. 1. EVs charging load modeling is the basis for further EMS applications. EV is different from any kind of conventional load demands because of its spatial movable feature. So in our work on modeling an EV or EV fleets, we utilize the traffic information from the transportation system — not only the
2
hisstorical statistics informattion, but also the real-tiime infformation. W With the estab blished modell, security asssessment can be peerformed to heelp the operato ors evaluate iff the grid is with w ennough margin when w large-scale EVs integrrated. In orderr to asssess the futuree influence, thee real-time trafffic information n is neecessary to pred dict the possiblle power injecttions to the grid d. A After security assessment, two possible results may be deetermined: norm mal or emergen nt. If the powerr grid is emerg gent, whhich means it i is not with h enough maargin during the folllowing short time or may y be danger iff some speciffied coontingency hap ppens, correctiv ve control or preventive p control haas to be carried d out to ensure the operating security. s The EVs E chharging facility y may be a new w alternative emergency control meethod. In contrrast, if it is seccure enough, how to involve the EV Vs into the po ower system optimized o opeeration is anotther intteresting probllem. Especiallly, thanks to the t EVs’ storrage chharacteristics, it is expectted to be coordinated and a coomplemented with w the interm mittent renewable sources and a finnally to reducee the carbon emissions of th he total electriccity prooduction. E EVs act as thee bridge betweeen the powerr system and the traansportation sy ystem. Currentt works mostly y focused on the inffluence to the power grid, while, w it shoulld not be igno ored thaat the power sy ystem itself maay also take po ositive or negattive efffects on the trransportation system if the number n of EVss is larrge enough. Fo or example, congestion on po ower transmisssion proobably leads to o traffic jams since s some chaarging services are noot available. So o information feedback f from m power system m to traansportation sy ystem is neceessary to realizze the optimized opperation for thee both two systeems.
behavioors are closelyy related to thhe drivers’ haabits. So the difficultty in EVs load model is how to deal with thhe stochastic characteers, not only teemporal but alsso spatial. Accordding to [8], thhe total charging load follow ws a normal distributtion if the EV Vs swarm is laarge enough, so the most importan ant two parameeters are μ t aand σ t2 , whichh denote the mean annd variance off the distributioon respectivelyy. In order to fully coonsider the EV Vs’ dependencyy on the driverrs’ behavior, we are trying to discover EV looad model froom massive HTS (Nationall Household historicaal travel data, such as NH Travel Survey) statiistics database released byy the U.S. ment of Transpportation, whicch records a ggreat amount Departm of travell information [[9]. First w we convert thee driving behaavior of conveentional cars into thoose of the EVs and generate lots of samplees with some importan ant factors as innput. Some tyypical factors too be studied include the batteryy capacity, the charging rate of infrastruuctures, the chharging profiless of the ownerrs and so on. A multipple regression method is adoopted to generaate a concise linear m model, with which the charging loadd curve is determinned as soon aas the factors are input. Moreover, the impacts of different ffactors on thee charging load curve can This proposed m model method is actually a also be iinvestigated. T kind of data mining. T The outline of modeling is shhown in Fig. 2.
Fig. 22. EVs charginng load modeling based on naaturalistic travel patterrns
Fig. 1. The Framework F off EMS with EV Vs Integration III. CURRENTT RESEARCHES A. Modeling E EVs take effeect on the po ower grid only after they are coonnected and charged, c so th he charge (inccluding dischaarge he premise of energy manag gement. Since the alsso) model is th im mpact on the grid is not sign nificant unless the EVs scalee is larrge enough, th he load modeel research is more concern ned abbout EVs’ aggregator mod del rather thaan a single EV. E ds, EVs charg ging Diifferent from other kind off load demand
B. Asseessment The im mpacts on thee power grid ccaused by EVss integration mainly include the ooverloading of transformerss and lines, voltage sags, transmisssion losses, phhase imbalancee, harmonics and so on. There havve been manyy works on thee evaluation methodss [10-12]. In tthis section, fuurther discussioon is mainly about tw wo issues. The ffirst one is thhe load forecaasting, withoutt which the contingeency assessmeent cannot bee made compprehensively. EVs aree movable andd may be connnected into thee power grid anytimee and anywherre. So it is cruccial to measurre the spatial informaation of EVs annd make moree precise forecaast based on the posssibility of connnecting spot aand charging rrequirement. ITS (Int ntelligent Transsportation Sysstem) is suppoosed to send this kindd of real-time traffic informaation to the eleectric power control center, such aas GPS coordinnates and veloocity of each EV. Appart from that,, the electricityy information of EVs are also inddispensable, suuch as the SOC C (State of Chaarge) and the battery capacity. Com mbining the traaffic informatiion with the
3
eleectricity inform mation of the EVs, E the meth hodology for lo oad forrecast of EVs can c be develop ped. T The second on ne is about thee power flow method, m which h is baasis for assessm ment. Currentlly, most researrches are focu used onn the distributio on grid and make m an equivaalent model of the traansmission grid d, PV nodes fo or root buses, for example. The T poower flow com mputation witho out considering g the transmisssion griid response maay result in inaaccurate resultss, especially wh hen thee EVs loads are large enough h. Thereby, an algorithm of GPF G (gllobal power flo ow) is developed which makees iteration on the boorder of the tran nsmission systeem (TS) and diistribution systtem (D DS). Through th he border information exchaange, both TS and a DS S can consider the response of o each other and a ensure a more m preecise result. Baased on GPF, some s assessmeent will be carrried ouut such as voltage v sag and a thermal problems. The T arcchitecture of GPF assessm ment of EV Vs integration is preesented in Fig. 3.
load deemands. In fuurther, all the reference vaalues of the threshollds may be cooordinated andd refreshed by the systemlevel coontrol.
Control Cen nter
Set Thresho old
Adjusting the t Charging Po ower of Chargin ng Facilitiess
Underfrequncy y Load Shedding g Coordination n
Undervoltagee Load Sheddin ng
Fig. 44. The Flowchaart of the Emerrgency Controll with EVs
Fig. 3. Thee Architecture of GPF Assesssment of EVs gration Integ C. Emergency Control C T To protect the power system m from unusuall disturbance, the schhemes of underfrequency lo oad shedding and undervolttage loaad shedding have h been dessigned [13,14]]. However, lo oad shedding will grreatly influencee the consumers with inevitaable loaad losses. If laarge-scale EVs are integrated d and the charg ging loaad takes an esssential part of the overall dem mands, there may m bee a new method dology to keep the grid secure by adjusting the chharging power instead i of direcctly shedding the t normal load ds. G Generally, the charging faccilities are con ntrollable on the chharging rate. Considering thaat the vehicles park in the most m tim me of the day and a the penetraation level of EVs E is high in the futture, EVs can provide a con nsiderable regu ulation capabillity. Esspecially in Ch hina, the Statee Grid and thee China South hern Poower Grid prom mote the battery y swapping as the major way y to suppply the electtrical energy to t EVs. The swapped s batterries wiill be collected d and charged d together, wh hich act as acttive loaads and can be directly contro olled by the griid. A threshold of frequency or o voltage wiill be set to the chharging facilitiees. When the power system iss under frequen ncy off under voltag ge, a local control c will be automaticaally triggered to redu uce the charging rate. The flowchart of the coontrol is Fig. 4. This round of o emergency control c should d be cooordinated with h conventionall load sheddin ng set-values and a takke action firstly to avoid in nfluence the cu ustomers’ norm mal
D. Optiimization Whethher the EVs arre "low carbonn" is still an oppen question in the inndustry and ressearch fields. E EV is zero carbbon emission when drriving, but its eenergy comes ffrom the poweer system. So an EV should not be regarded as llow carbon unnless we use more annd more clean energy to chaarge it. On thee other hand, the mosst promising cllean energy is w wind power. H However, the inherentt intermittent nnature of wind power leads too significant challengge with the active powerr balance andd frequency control. It is notable that the large-sccale wind pow wer in the generatiion side and the large-scalee electric vehhicles in the demandd side are not oonly a severe cchallenge to thee future grid scheduliing, but also a great opportuunity. Wind poower and the electric vehicle, one in the energyy generation sside and the other iin energy coonsumption siide, have soome natural complem ment characterristic. For exam mple, wind poower appears to be reeverse-regulatioon that means generating moore power in f-peak demandd periods succh as midnigght. If EVs the offcoordinaated charge is implemented and absorb moore power at this tim me, it is positiive to make uuse of more w wind powers without abandon. In aaddition, the bbattery swap m mode as well as the rretired batteriees recycling caan be adopted as effective energy storages, whicch is the most effective wayy to stabilize the interrmittent wind ppower. The eleements of the ccoordination are show wn in Fig. 5.
4
Transmission System
Coord dination
Distribution System
Fig. 6. T The Architectuure of SCGS IV. CONCLUSSION Fig g. 5. The Wind d-EV Coordinattion A An optimizatio on model has been b proposed with wind pow wer abbandon minimization objecttive or emissiion minimizattion obbjective, consid dering carbon capture c powerr plants as welll as EV V aggregators [4]. The simulaation results sh how that the EV Vs’ chharging (and discharging) d beehaviors can be b complemen nted wiith the win nd powers. Thanks to the Wind--EV coomplementation n, both the ab bandoned wind d power and the ovverall carbon em missions are grreatly reduced. E. Feedback witth ITS E Electric vehiclles couple the power grid and a transportattion neetwork closely which both arre complex maan-made system ms. Feew interactionss of the two systems are considered no ow, esppecially the in nfluence to the transportatio on system by the poower grid. Forr example, thee existing on--board navigattion system may guide the EV to o the nearest charging stattion whhich perhaps iss under power congestion unfortunately, wh hile at the same time, other chargin ng stations are idle. Therefore, a sm mart charging guide g system (S SCGS) is need ded to prevent this t sittuation from happening h whicch considering g both traffic and a poower operation information [7 7]. Thhe SCGS build a link betw ween the trafffic control cen nter (T TCC) and the power p system control center (PSCC) [7]. The T arcchitecture of SCGS S is preseented in Fig. . If an EV neeeds chharging in a trip p, the SCGS will w send its bassic information n to thee TCC. Utiliziing the techno ologies of ITS S, such as Glo obal Poositioning Sysstem, Geograp phic Information System and a W Wireless Comm munication, thee SCGS makees charging lo oad forrecast and send ds the result off forecast to thee PSCC. Then the PS SCC can perfo orm security asssessment for the grid with the chharging load forecast. fo A seccurity index for fo each charg ging staation is generatted, which is used u for revisin ng the geograp phic disstances of the path of the triip. The revised d distance can n be called as the overall distance.. The overall distance not only o refflects the geog graphic distance of the path, but b also consid ders thee security of the t grid. In sh hort, the overaall distance is the baalance between n the traffic dem mand of the EV V’s owner and the opperation deman nd of the grid.. The charging g station with the shortest overall distance is reccommended to o the owner. As A a ressult of the feeedback from the power grrid, the poten ntial neegative impact of fast chargiing in charging g stations willl be miitigated.
Some preliminaryy works onn energy m management technoloogies considerring EVs integgration have bbeen briefly introducced in this papeer, most of whhich are ongoinng now. Four commonn conclusions sshould be empphasized: 1) Thee highlight of our research is based on thee conception of tthe system of ssystems, whichh involves bothh the electric pow wer system andd the transportaation system. 2) Thee interactions oof both inform mation and enerrgy between EV Vs and power ggrids should bbe supported bby the EMS withh EVs integrattion. 3) Cooordination beetween EVs annd renewable energies is esseential for the bboth, which willl play an impoortant role to reduuce the carbonn emission. 4) Thee distributionn grid connnects tightly with the trannsmission grid,, so the global viewpoint is nnecessary. V. REFERENC CES [1]
Pike Research E Electric Vehicle M Market Forecasts S Study [Online]. Available: http ://inhabitat.com/5--2-million-strong-report-predicts46x-increase-in--ev-sales-by-2017//rsz-pike-research--ev-salespredictions-chart rt/. [2] W. Kempton annd J. Tomic, "Vehhicle-to-grid powerr fundamentals: Calculating capaacity and net revenue," Journal of P Power Sources, vol. 144, pp. 2688-279, 2005. [3] C. Guille and G G. Gross, "A conceeptual framework for the vehicleto-grid (V2G) im mplementation," E Energy Policy, voll. 37, pp. 43794390, 2009. Z. Li, H. Sun, Q [4] Q. Guo, Y. Wang,, and B. Zhang, "S Study on windEV complemenntation in transmission grid side," in Power and Energy Society G General Meeting, 22011 IEEE, 2011, pp. 1-12. [5] J. A. Pecas Lopees, P. M. Rocha A Almeida and F. J. Soares, "Using vehicle-to-grid to maximize thhe integration oof intermittent renewable energgy resources in isslanded electric grids," in Clean Electrical Poweer, 2009 Internatiional Conference on, 2009, pp. 290-295. C. D. White andd K. M. Zhang, "U Using vehicle-to-ggrid technology [6] for frequency rregulation and peeak-load reduction," Journal of Power Sources, vol. 196, pp. 39722-3980, 2011. Q. Guo, Y. Waang, H. Sun, Z. L Li, and B. Zhang, "Research on [7] architecture of IITS based Smart Charging Guide System," IEEE PES General Meeeting, Detroit, Jully 24th –July 28th, 22011 D. Wu, D. C. Aliprantis and K K. Gkritza, "Electrric Energy and [8] Power Consumpption by Light-D Duty Plug-In Elecctric Vehicles," Power Systems, IEEE Transactionns on, vol. 26, pp. 738-746, 201101-01 2011. U.S. Departm [9] ment of Transsportation, Fedeeral Highway Administration, 2009 National H Household Travell Survey. URL: http://nhts.ornl.ggov. [10] S. Rahman and G. B. Shrestha, "An investigation into the impact of electric vehiccle load on the eleectric utility distribbution system," Power Delivery, IEEE Transactioons on, vol. 8, pp. 591-597, 199301-01 1993. [11] P. Richardson, D. Flynn and A.. Keane, "Impact assessment of varying penetrrations of electrric vehicles on low voltage
5 distribution systems," in Power and Energy Society General Meeting, 2010 IEEE, 2010, pp. 1-6. [12] M. Etezadi-Amoli, K. Choma and J. Stefani, "Rapid-Charge Electric-Vehicle Stations," Power Delivery, IEEE Transactions on, vol. 25, pp. 1883-1887, 2010. [13] C. Concordia, L. H. Fink and G. Poullikkas, "Load shedding on an isolated system," Power Systems, IEEE Transactions on, vol. 10, pp. 1467-1472, 1995-01-01 1995. [14] C. W. Taylor, "Concepts of undervoltage load shedding for voltage stability," Power Delivery, IEEE Transactions on, vol. 7, pp. 480488, 1992-01-01 1992.
VI. BIOGRAPHIES Qinglai Guo (M’2009) was born in Jilin City, Jilin Province in China on Mar. 6, 1979. He graduated from the Department of Electrical Engineering, Tsinghua University, Beijing, China, in 2000 with B.S. degree. He received his PhD degree from Tsinghua University in 2005 where he is now an associate professor. He is member of CIGRE C2.13 Task Force on Voltage/Var support in System Operations. His special fields of interest include the EMS advanced applications, especially the automatic voltage control and V2G. Hongbin Sun (M’2000) received his double B.S.degrees from Tsinghua University in 1992, the Ph.D from Dept. of E.E., Tsinghua University in 1997. He is now a full professor in Dept. of E.E., Tsinghua Univ, and assistant director of State Key Laboratory of Power Systems in China. From 2007.9 to 2008.9, he was a visiting professor with School of EECS at the Washington State University in Pullman. He is member of IEEE PES CAMS Cascading Failure Task Force and CIGRE C2.13 Task Force on Voltage/Var support in System Operations. His research interests include energy management system, voltage optimization and control, applications of information theory and intelligent technology in power systems. E-mail:
[email protected] Yao Wang received his Bachelor degree from the School of Electrical Engineering at Northeast Electric Power University in 2010 and now he is pursuing the Master degree in the Department of Electrical Engineering at Tsinghua University. His research interests include the charging of electric vehicles and V2G.
6
Current Energy Management Technologies Research in China Considering EVs Integration Response to Reviewers Qinglai Guo, Member, IEEE, Hongbin Sun*, Member, IEEE, Yao Wang, Zhengshuo Li, Boming Zhang, Fellow, IEEE
Dear Editor, dear Reviewers: We thank you for your comments. It is very helpful in improving our manuscript. The followings are our responses. Responses to Reviewer 1: Comments: “Conclusion is needed.” Response: Thank you for your suggestion. Actually, the last section should be “Conclusion”. However, it is noted as “Summary” in the original version. Now we have renamed this section as “Conclusion”. Furthermore, we have emphasized the four conclusions related to our preliminary work. After modification, the “Conclusion” section is as the following: “ Some preliminary works on energy management technologies considering EVs integration have been briefly introduced in this paper, most of which are ongoing now. Four common conclusions should be emphasized: 1) The highlight of our research is based on the conception of the system of systems, which involves both the electric power system and the transportation system. 2) The interactions of both information and energy between EVs and power grids should be supported by the EMS with EVs integration. 3) Coordination between EVs and renewable energies is essential for the both, which will play an important role to reduce the carbon emission. 4) The distribution grid connects tightly with the transmission grid, so the global viewpoint is necessary. ” Again, we are grateful for your help on improving the paper.
