Iteratioal Coferece o Eergy ad Evirometal Protectio (ICEEP 206) Operatioal Status Evaluatio for Electric Vehicle Chargers based o Layered Radar Map Method Tao Jiag,a, Weiyog u,b, Jia Hu, c, Zhe Luo,d, ushuag Li2,e ad Fa e2,f uhag Power Supply Compay of State Grid Zheiag Hagzhou Electric Power Compay, Hagzhou 300, Chia; 2 State Key Laboratory of Power Trasmissio Equipmet & System Security ad ew Techology, Chogqig Uiversity, Chogqig 400044, Chia a 45397022@qq.com, b7453288@qq.com, c87965882@qq.com,d529737265@qq.com, e 45592205@qq.com Keywords: electric vehicle chagers; status evaluatio; idices selectio; factor aalysis; radar map Abstract. For optimizig the maiteace strategy of electric vehicle (EV) chagers, it s ecessary to have a comprehesive evaluatio for it by usig its moitorig data. From the battery, electrical machiery trasmissio ad charge-discharge process, 26 characteristic quatities are chose as EV chagers operatio state idicators, the commo factor idicators are extracted by usig the factor aalysis method ad the the EV chagers state idicators are classified by the idepedet commo factors; the classified idicators are screeed by usig stepwise regressio aalysis, ad the idicators that cotribute little to the evaluatio result are excluded, thereby the EV chagers operatio state comprehesive evaluatio system are established. Combied hierarchical structure of EV chagers evaluatio ad radar map method, the ed radar map method is proposed for EV chagers comprehesive evaluatio. Itroductio Compared to covetioal cars, electric vehicles have icomparable advatages i promotig eergy coservatio ad emissio reductio, reducig operatig costs ad geeralizig ew eergy utilizatio []. With EV abudatly available, it will have ifluece o the security ad stability of power gird, ad also brig hardship to the repair ad maiteace of EV chargers [2-4]. Moitorig ad evaluatig the operatig state of EV chargers timely ad accurately, ca help deal with the fault ad optimize the maiteace strategy of EV chargers. The EV chagers operatio state comprehesive evaluatio system based o visualizatio techology, ot oly ca visualize the relatioship amog the idicators, but also guide the repair strategy graphically, which must solve the classificatio ad the reductio of dimesio of multiple idicators, ad the graphical display of assessmet results [5]. Accordig to the battery, electrical machiery trasmissio ad charge-discharge process, this paper chooses EV chagers operatio state idicators comprehesively ad obectively. The EV chagers state idicators are classified by usig the factor aalysis method ad the idicators that cotribute little to the evaluatio result are excluded by usig stepwise regressio aalysis, thereby the EV chagers operatio state comprehesive evaluatio system are established. The ed radar map method is proposed for EV chagers comprehesive evaluatio makig the differet levels of EV chagers multi-dimesioal status idicators represet i the radar map visually. Selectio of EV chagers operatio state idicators The electrical structure of the EV chagers icludes the power coverter, storage battery ad electrical machiery. Accordig to the battery ad electromotor state characteristic quatities i the ruig process, ad the electrical characteristic quatities of the charge-discharge process, 26 idicators are chose as EV chagers operatio state idicators, as show i Table. 206. The authors - Published by Atlatis Press 96
Tab. Operatioal status idices of electric vehicles o ame o ame Chargig voltage deviatio 4 Battery termiatio voltage 2 Chargig frequecy deviatio 5 Discharge duratio 3 Chargig voltage flicker 6 Battery output 4 Chargig voltage ubalace 7 Self-discharge rate 5 Chargig voltage fluctuatio 8 Battery cycle life factor 6 Chargig power factor 9 Battery temperature matchig coefficiet 7 Chargig curret harmoics 20 The iitial SOC matchig coefficiet 8 Eergy loss 2 Battery voltage matchig coefficiet 9 Discharge efficiecy 22 Motor temperature 0 Eergy cosumptio per mile 23 MTBF Discharge depth 24 Maximum speed 2 Charge ad discharge times 25 Drive distace 3 Battery capacity 26 Motor overload factor The operatio state of the battery, electrical machiery ad other differet modules ca be obtaied by usig the state idicators, ad the overall performace of EV chagers ca also be obtaied through a comprehesive evaluatio. Because of the correlatios betwee the differet modules ad the differet idicators i the same module of EV chagers, it will have a strog impact o the evaluatio result. So, there is a eed to classify ad scree the EV chagers state idicators that give i Table. Classificatio ad screeig of EV chagers operatio state idicators based o factor aalysis Assumig the umber of samples EV chagers operatio state is ad the umber of idicators is m=26, where V ad V 2 represet reverse ad appropriate idicators set respectively. The -th idicators of the i-th sample treds ito, max{ x } x, v V i x () max{ x x } x x, v V2 i Where x " ad x ' is the idicators before ad after the tred, x " is the appropriate of the -th idicators (appropriate idicators x " closer to x " is better). The calculatig the stadardized state idicators x, x x i x 2 ( x x ) i i The commo factor state idicators are extracted by usig the factor aalysis method to the sample data, which will cotai multiple state idicators with close relatioship. Ad the the EV chagers state idicators are classified whereby it. (2) 97
Eter the stadardized data of samples ad m idicators Seek sample correlatio matrix R ad characteristic, rakig l >l 2 > >l m >0 Eter X i ad y i of the i-th classificatio From the regressio equatio oly cotaiig costat term, calculate the cotributio rate F of each idicators Select the first k idicators, so the cumulative cotributio rate of k m characteristic li li > 0.85 i i Seek eigevectors of k eiges calculate factor loadig matrix A Without itroducig idicators vmax F max >F i? D>Tol? Itroduce vmax to the idicators set F mi <F out? Without Excludig idicators vmi Update idicators set Vc after screeig Exclude idicators vmi Categories whether explicit physical meaig? Factor rotatio A'=AG Accordig to V c, calculate the cotributio rate F of each idicators Make idicators a' >a t ito the -th classificatio Output the classificatio result of m states idicators of EV chagers F (V c )>F i 且 F (V c )<F out? Output the optimal idicators set V c Fig. Flow chart of state idicators classificatio Fig. 2 Flow chart of status idices selectio Flow chart of state idicators classificatio by factor aalysis is show i Fig.. Accordig to the real-time moitorig data of a EV chager, the coefficiet matrix characteristic related to the sample data ca be obtaied whe state idicators data are stadardized. Ad the 26 EV chagers operatio state idicators come dow to 5 types of f to f 5, (commo factors), icludig battery efficiecy, battery lifetime, battery charge-discharge characteristic, battery cell cosistecy ad electrical machiery operatio state. Where battery efficiecy cotais 3 idicators of V 8 to V 0, battery lifetime cotais 5 idicators of V to V 2 ad V 7 to V 8, battery charge-discharge cotais 9 idicators of V to V 7 ad V 4 to V 5, battery cell cosistecy cotais 3 idicators of V 9 to V 2, electrical machiery operatio state cotais 6 idicators of V 9, ad V 22 to V 26. Choosig a skimp umber of idicators will icrease the evaluatio error of EV chagers operatio state, however, a excessive umber of idicators will icrease the complexity of models ad reduces its practicability. Thus, it is ecessary to scree the 5 subsets of classificatio idicator of f to f 5 further. Flow chart of state idicators selectio by stepwise regressio aalysis is show i Fig. 2. Calculatig the cotributio rate of the idicators F, i accordace with the i-th classificatio idicators set V c after stepwise selectios. If F max of the max -th idicator satisfies F max =max(f )>F i ad acceptable level D>Tol, V c itroduces it, otherwise ot. If F mi of the mi -th idicator satisfies F mi = mi(f )<F out, V c excludes it, otherwise ot. Whe all the cotributio rate of the idicators i the V c is greater tha F out, ad outside the V c is less tha F i, the Screeig eds ad the optimal idicators set ca be outputted. The EV chagers operatio state comprehesive evaluatio system are established by the selectio, classificatio ad screeig of EV chagers operatio state idicators, as show i Fig. 3. The figure icludes 3 s of a target, a classificatio ad a idicator. The target is divided ito 5 categories, ad refies the idicators uder each category. For easy aalysis, the idicators uder differet categories are reumbered. 98
target classificatio battery efficiecy f idicator eergy loss f discharge efficiecy f2 eergy cosumptio per mile f3 discharge depth f2 evaluatio of EV chargers operatio state f battery lifetime f2 battery chargedischarge characteristic f3 battery cell cosistecy f4 electrical machiery operatio state f5 battery cycle life facto f22 charge ad discharge times f23 battery capacityf24 chargig voltage ubalace f3 chargig voltage fluctuatiof32 chargig power factorf33 chargig curret harmoics f34 chargig curret harmoics f35 discharge duratio f36 battery temperature matchig coefficietf4 the iitial SOC matchig coefficiet f42 battery voltage matchig coefficiet f43 motor temperature f5 MTBF f52 maximum speed f53 drive distace f54 motor overload factor f55 Fig. 3 Evaluatio idices system for operatioal Comprehesive evaluatio for EV chargers operatio state based o ed radar map method The EV chagers operatio state comprehesive evaluatio ca be divided ito three- aalysis models by the method described above. Combied the hierarchy structure i Fig. 3 with the radar map method, the ed evaluatio method of EV chagers operatio state ca be achieved. Schematic diagram of comprehesive evaluatio based o ed radar map is show i Fig. 4. First, the radar map of each ca be obtaied usig the sample data of the idicator, ad got the radar map evaluatio of each classificatio through the image feature. Ad the accordig to the evaluatio, the evaluatio result of the target of the EV chagers operatio state ca be obtaied. The assessmet steps of ed radar map method are as follows. ) Calculate the comprehesive s of the idicator ad classificatio combied with the improved subective ig method ad obective etropy method. 2) Draw the first uit legth lie OA vertically up, takig the ceter as the startig poit. Lettig the first idicator covert to a agle, draw the secod uit legth lie OB, ad the rest lies are i the same way. 3) Draw the bisector of agle of each sector, takig the ceter as the startig poit. Each ormalized idicator is used as the legth of the bisector of agle, ad coectig the periphery poits sequetially to get the radar map. Usig the total area of polygo S ad the perimeter C, the quatitative evaluatio SC is obtaied. 4) Repeat step 2) ad 3) to get the radar map ad its evaluatio of 5 categories. 5) Accordig to the evaluatio of 5 categories, the radar map ad comprehesive evaluatio of the EV chagers operatio state is obtaied by step 2) ad 3). Ad the comprehesive level of the sample waitig for the evaluatio is determied by the differet levels of stadard samples. 99
The comprehesive evaluatio result of EV chargers operatio state radar map evaluatio idicator target battery efficiecy evaluatio idicator radar map evaluatio liear Weight idicator battery lifetime evaluatio idicator radar map evaluatio liear Weight idicator battery cell cosistecy evaluatio idicator classificatio radar map liear evaluatio Weight idicator eergy loss discharge efficiecy eergy cosumptio per mile discharge depth battery cycle life factor charge ad discharge times Battery capacity battery temperature matchig coefficiet the iitial SOC matchig coefficiet battery voltage matchig coefficiet idicator Fig. 4 Schematic diagram of comprehesive evaluatio based o ed radar map Coclusios This paper divided the 26 EV chagers operatio state idicators ito five classificatios, icludig battery efficiecy, battery lifetime, battery charge-discharge characteristic, battery cell cosistecy ad electrical machiery operatio state by factor aalysis. Ad the redudat idicators of each classificatio are screeed by usig stepwise regressio aalysis, thereby the practical EV chagers operatio state comprehesive evaluatio system are established. Combied hierarchical structure of EV chagers evaluatio ad radar map method, the comprehesive evaluatio result of EV chargers operatio state is obtaied. The EV chagers operatio state comprehesive evaluatio system proposed i the paper ca reflect the real state of EV chargers better, ad the evaluatio result ca provide techical referece for the maiteace of EV chargers. Refereces [] Wag Xifa,Shao Chegcheg,Wag Xiuli,et al.survey of electric vehicle chargig load ad dispatch cotrol strategies[j].proceedigs of the CSEE,203,33():-0. [2] ilmaz M,Krei P T.Review of battery charger topologies, chargig power levels, ad ifrastructure for plug-i electric ad hybrid vehicles[j].ieee Trasactios o Power Electroics, 203,28(5):25-269. [3] Li Huilig, Bai Xiaomi. Impacts of electric vehicles chargig o distributio grid [J].Automatio of Electric Power Systems,20,35(7):38-43. [4] Zhou iacheg,xiog Xicog,Wag Qiaggag.Simulatio of chargig load probability for coectio of differet electric vehicles to distributio etwork[j].electric Power Automatio Equipmet,204,32(2):-7. [5] Wag Deqig,Wa ogbo,wag Xiag,et al.improved radar chart based o pricipal compoet ad its applicatio i comprehesive evaluatig [J].Joural of Applied Statistics ad Maagemet,200,29(5):883-889 200