Set up and test of a LiFePO4 battery bank for electric vehicle
Contents
1. April 2011 at www jun si com UploadFiles CellLog_8S pdf 7 Alekssander Santiago Paulo G Pereirinha Joao P Trov o Charge and Discharge Test Bench Monitoring System for Lithium lon Batteries for Electric Vehicle International Youth Conference on Energetics Leiria 2011 8 LogView software www logview info vBulletin 9 Muhammad H Rashid Power Electronics Circuits Devices and Applications Second Edition 1993 10 J Trov o F Ferreira Distor o Harmonica no Sector Industrial Causas Efeitos e Solu es Revista Manuten o N 88 Janeiro Mar o de 2006 11 Jo o P Trov o Frederico M Santos Marco J Silva Humberto M Jorge A Web Based Monitoring Approach for Power Systems in Industrial Plants Proc of IEEE International Symposium on Industrial Electronics ISIE 08 Cambridge UK 30th June 2nd of July 2008 Authors prof dr Paulo G Pereirinha IPC ISEC Rua Pedro Nunes 3030 199 Coimbra Portugal E mail ppereiri isec pt MSc Eng Joao P Trovao idem E mail jtrovao isec pt BSc Aleks Santiago idem E mail agsantiago120 gmail com PRZEGL D ELEKTROTECHNICZNY Electrical Review ISSN 0033 2097 R 88 NR 1a 2012 1972. BMS when the cells reach a certain level of charge the voltage starts to rise exponentially This area of the curve is a very unstable zone since the voltage changes very fast requiring special care to not exceed the maximum recommended voltage for the cells With recourse to CellLog devices and the interruption system this problem can be prevented by taking an interruption when some cells reach the level of voltage near the maximum After the interruption the voltage level falls rapidly to a value much lower than it was previously This value is not a fixed value varies from cell to cell but is close to 3 4 V 2 The chargers Assesses the level of battery SOC and adapt the charge currents This kind of charge has the ability to maintain a constant current for SOC steps thereby optimizing the load In terms of energy quality for smaller currents this causes a harmonic distortion that is within the limits allowed about 3 for the voltage THD and 111 for current THD However with such chargers generalization some problems might arise in the distribution network Regarding the efficiency it is reasonable and corresponds to the manufacturer prediction approximately 85 The values of THDi and performance vary when varying the current and when the THDi increases the efficiency decreases Indeed a current reduction causes a higher increase in the amplitude of the harmonics than in the fundamental frequency raising their influence
3. a real EV utilization an embedded system and with cell balancing capacity is need To fulfil this purpose it was used a Real Time Battery Management System RI BMS from GWL Power which is a 196 70 management system applicable to any kind of lithium battery This acts during the charge and discharge of the batteries controlling individually each cell through Cell Balancing Units CBUs as in Fig 14 that communicate with a Master Unit Fig 15 through a flat cable FC 10 The Master Unit can work standalone or connected to the computer which through software supplies all the information about currents and voltages of each cell and of the battery bank Its bigger advantage resides in that unlike the Cel Log 8S beyond monitoring also controls the processes It is capable of balancing the voltage of each individual cell according to the definitions of the user previously stored into the Master Unit through the computer When during the charge process one cell reaches a user defined voltage its CBU starts deviating up to 5 A of the charge current allowing this cell to charge slower than the others and consequently to prevent its overcharge while the other cells are increasing their charge The RT BMS also possesses alarms for anomalies that can occur such as errors or lowering of the voltages for risky levels Fig 15 Master RT BMS Control Unit case open and CellLog 8S Before starting the tests with the RT BMS instal
4. bank voltage at 15 A discharge a 2000 1800 1600 1400 1200 1000 Discharge Power W Discharge Power for 15 A In Fig 12 it is presented the evolution of the individual cells and accumulated blue higher curve voltages of the 8 cells monitored by one CellLog and shown in LogView during a charge followed by a discharge As expected from Fig 1 at the beginning and at the end of both charge and discharge the voltage variations are more pronounced It has also been verified that this effect increases with the increase of the discharge current In Fig 12 the referred phenomenon is more clearly seen particularly at the end of the charge where some of the individual cells voltage suddenly increases very sharply and decreases exponentially at the beginning and at the end of the discharge Due to this variation which do not occur in all the cells at the same time is it mandatory the existence of an individual cells monitoring system with an interruption system to avoid the damage of the cells by reaching very high or very low voltages 195 Sh 33m 20s 6h S6m 40s 8h 20m 00s 4h 10m 00s 0 1h 23m 20s 2h 46m 40s Fig 12 8 cells voltage variation total and individual voltages The results for the diferent currents were Cycle of 15 A discharge current Energy Absorbed at charge 7393 02 W h Energy Provided at discharge 7226 25 W h Cycle Efficiency 97 74 Cy
5. distortion rates showing and recording their values The LabVIEW system was built using the acquisition board PCI MIO 16E 4 of National Instruments with the I O connector block SCB 68 that enables the use of 16 non differential channels or 8 differentials The channels are connected to insulated current and differential voltage probes 7 The battery pack efficiency is determined dividing the energy provided during its discharge by the energy absorbed in the charge The charger efficiency is calculated dividing the energy supplied to the batteries during a charge operation by the energy it absorbs from the network The charger instantaneous efficiency is determined dividing the instantaneous power absorbed by the batteries by the instantaneous power absorbed through the grid The total current harmonic distortion THDi of the current chargers is calculated instantaneously The input power P 1 and the input energy E 3 absorbed by the charger are calculated by 1 P V I FP where FP is the power factor and in non sinusoidal system can be calculated by 2 9 11 2 rpp 9 _ 1 THD 3 E Pat As mentioned before the bank configuration is different for the charge a series of two groups with a series of 16 cells and the discharge 32 cells in series Therefore there was need to develop two monitoring programs one for charge Fig 5 and one for discharge 7 Its visual interface has several scopes graphic d
6. Paulo G PEREIRINHA Jodo P TROVAO Alekssander SANTIAGO Polytechnic Inst of Coimbra 1 Inst for Systems and Computers Engineering at Coimbra 2 Portuguese Electric Vehicle Association 3 Set up and test of a LiFePO4 battery bank for electric vehicle Abstract This paper deals with the set up and characterization tests of a lithium iron phosphate LiFePO4 battery bank for an electric vehicle he first charge and discharge a simple system to monitor and ensure tight voltage limits and a LabVIEW system to measure the voltages currents and calculate the battery pack effective capacity and charge discharge efficiency are presented The efficiency and harmonic contents of the used chargers are measured The installation of a commercial battery management system and its impact in previous results are also discussed Streszczenie W artykule przedstawiono testy parametr w banku baterii lithium iron phosphate LiFePO4 w zastosowaniu do pojazd w elektrycznych Przedstawiono system do testowania pojemno ci baterii oraz skuteczno ci adowania i roz adowania System do badania baterii LiFePO4 w zastosowaniu do pojazd w elektrycznych Keywords Battery bank LiFePO chargers setting up and experimental tests electric vehicles S owa kluczowe baterie samochodowe LiFePO uktady taDOWANIA BATERII Introduction Electric vehicles are expected to be a major contribution to sustainable mobility The energy storage capacity is
7. Thus in percentage terms the harmonics increase in amplitude 3 The battery pack Without the balancing system the charge and the discharge of batteries is not complete and either one or the other process was terminated as soon as any cell enter the unstable zone and hit the voltage limit established Thus the energy absorbed and supplied by the batteries did not reach its maximum value This is also reinforced as usually the cells only reach its maximum capacity after some cycles In these tests these values vary between 7200 Wh and 7700 Wh resulting in a maximum efficiency near of 98 and decreasing with the discharge current 4 The RT BMS Concerning overcharge prevention the RT BMS operation is based on diverting part of the charge current entering the cell as a kind of by pass and when necessary will discharge the higher charged cells That may lead to an increase of the energy consumption In terms of energy results it was quite identical to the tests without the balancing system due to the lack of balancing power from the balancing system This can be improved for example by a better CBUs heat removal As a final conclusion the study done allowed to improve the authors knowledge on the LiFePQO batteries and prepare a battery bank for EV Acknowledgment The authors want to thank the Miguel A C da Fonseca BSc in Electrical Engineering former ISEC student for his collaboration in the preparation of the b
8. age is 4 25 V and minimum discharge voltage is 2 5 V To build the used battery pack 32 cells were grouped It should be noted that the cell voltage varies both during charge and discharge Fig 1 but considering 3 V for the cell nominal voltages the pack has 96 V nominal voltage and C 90 Ah and an energy stored of 8640 Wh The chargers play an important role in the integration of electric vehicles since they recharge the batteries and must be adequate to the battery chemistry the levels of voltage and type of charge In this study it was used the single phase POW48V30A model from GWL Power company prepared to work with lithium ion batteries This model has intelligent control that calculates the battery pack state of charge SOC depending on its voltage and adjusts the current However this model can only charge 16 cells and should not be used with fewer cells Consequently two of these chargers were used except for the first charges and the battery pack had to be divided in 2 groups of 16 cells during the charging The nominal voltage is 48 V and maximum voltage 64 V none oes MUGS C a TABAY 2 5 V ah abl sou so Fig 1 Charge discharge curves from manufacturer manual 5 of TS LFP90AHA cells bottom left First charge and discharge Even though manufacturer s manual allowed higher currents the first cycle of charge and discharge was performed with low current 3 A The objective of this first cycle was
9. attery and some experimental tests This work has been partially supported by FCT under project grants PEst C EEI UI0308 2011 MIT SET 0018 2009 and MIT MCA 0066 2009 REFERENCES 1 P G Pereirinna J P Trov o A Marques J Silvestre F Santos A Campos M Silva P Tavares The Electric Vehicle VEIL Project A Modular Platform for Research and Education Proc of the 2 nd European Ele Drive Conference EET 2007 Brussels Belgium May 30 June 01 2007 2 Paulo G Pereirinha Joao P Trovao L Marques M Silva J Silvestre F Santos Advances in the Electric Vehicle Project VEIL Used as a Modular Platform for Research and Education Proc of the EVS24 International Battery Hybrid and Fuel Cell Electric Vehicle Symposium Stavanger Norway 13 16 May 2009 3 J P Trov o P Pereirinha H Jorge Design Methodology of Energy Storage Systems for a Small Electric Vehicle World Electric Vehicle Journal Volume 3 2009 ISSN 2032 6653 4 Marco Silva Joao P Trov o Paulo Pereirinha Luis Marques Multiple Energy Sources Monitoring System for Electric Vehicle Proc of the 19 th International Symposium on Power Electronics Electrical Drives Automation and Motion SPEEDAM 2008 Ischia Italy 11 13 June 2008 5 Thundersky Company presently Winston Battery Limited www thunder sky com home_en asp 6 CellLog 8S Cell Voltage Monitor amp Logger User s Manual Shenzhen Junsi Electronic Co Ltd available on 31
10. cle of 30 A discharge current Energy Absorbed at charge 7530 60 W h Energy Provided at discharge 7284 16 W h Cycle Efficiency 96 73 Cycle of 60 A discharge current Energy Absorbed at charge 7768 08 W h Energy Provided at discharge 6814 76 W h Cycle Efficiency 87 8 Sum of both groups of 16 cells From the above results even though if the energy absorbed by the battery pack is not equal for all tests it is clear that the energy provided at the discharge decreases with the increase of the discharge current That is due to the fact that the increasing of the discharge current increases the losses at the wires at the contacts and mainly the internal losses of the cells Higher currents also increase the voltage variation taking to an earlier end of the discharge preventing the maximization of the energy transferred at the process The measured battery pack efficiency is shown in Fig 13 As can be seen the efficiency decreases with the growth of discharge current and it is clearly shown why according to the manufacturers data sheet the standard discharge current is 0 3 C 27 A 100 Efficiency Rendimento 8 10 20 30 40 Discharge Current A Fig 13 Batteries Efficiency as a function of the discharge current Battery Management System Tests Results The LabVIEW developed system allows to evaluate the charger and battery pack efficiency and power quality Nevertheless for
11. d state relay contact is not galvanic isolated from the CellLog inputs having to be connected through auxiliary electromechanical relays 7 One has also to be very careful as the USB connector ground of each of the CellLogs is also directly connected to its cells lower voltage more than one Cel lLog cannot be connected to the same computer at the same time and even if different computers are used one for each Cel Log special care has to be taken to avoid short circuits through the different computers supplies or USB connectors ground loops The individual cells and total voltages and its time evolution can be visualized using the LogView software 8 During the first charges one important verified feature was that above 3 4 3 5 V the voltage on the cells increases very fast Even with a 3 A charge it takes just a few minutes to reach the maximum 4 0 4 25 V limit and when discharging bellow 3 1 V the voltage drops very fast cf Fig 12 for higher currents The first discharge was also done at 3A in order to gain Knowledge on the battery behavior System and Tests Methodology for Battery pack and Chargers Characterization After the first charge and discharge cycle to characterize the battery bank it was necessary to develop systems that allow monitoring and calculate the chosen parameters Thus a program was created in LabViIEW platform using an acquisition board to measure currents voltages and calculate power energy harmonic
12. ents and in Fig 8 the input and output powers of both chargers are presented 50 nm Fig 11 40 30 Currents A 20 Tens o Pack 1 V Corrente Saida Carregador 1 JA Corrente Saida Carregador 2 A 10 Group Voltage V Chargers Output Nuwouonrnvuouonroaowvowowodantrt tT uwuowWDonwTvuouwowowovel wo ow SAO TAMANMNMRNMHOMOAN DMI OAOMoaAnawrtoawTroa xw eooaow aorkt mawuoonnrordtnrwF te woHwm wonnananrvuoeanwoonrnr oo ot AA ANNN MOM TtT TT OT NO NNnWOHOUORR WWE A Time s 9800 Fig 7 16 Cells Group Voltage blue Charger 1 Output Current red and Charger 2 Output Current green After each charge the battery pack was prepared for the discharge test with the 32 cells in series as in Fig 9 The results in Fig 10 and Fig 11 are fora 15 A discharge current PRZEGL D ELEKTROTECHNICZNY Electrical Review ISSN 0033 2097 R 88 NR 1a 2012 0 wnon omar an NANO OD nnwoo 6 346 ON oO ct oO o a 1386 9098 9448 9800 Fig 8 Charger 1 Input Power blue Charger 1 Output Power green Charger 2 Input Power red Charger 2 Output Power purple e ETER ai aa fi a ta teas a Po on abs a ban sy pt F4 A i H Di o ao jc aid Fig 9 Battery bank connected for discharge test 32 cells in series Bench Voltage V Tens o do Pack V Fig 10 Battery
13. ger when the current has an abrupt decrease The charge powers Fig 4 have the same behaviour that the current showing a small difference at the charge end where it seems to increase faster This is due to the behavior presented by the cells voltage at the end of the charges as will be shown in Fig 12 Cahrge Current A Time s Fig 3 Charger Output Current 2000 1500 4 1000 4 Input Power W 500 4 P Entrada C P Sa da C a ee eee A A TOTTI OO N wotrowon co mo DD WMO Aa MO DW YM WN CO WOMM S H MNA MK CO DOD NN St Fig 4 Charger Input Power blue and Output Power red The evolution of the THDi and charger efficiency in Fig 5 is in accordance with the expected results At the moment of transition of stage both THDi and efficiency show a variation PRZEGL D ELEKTROTECHNICZNY Electrical Review ISSN 0033 2097 R 88 NR 1a 2012 Input Current THD an O x THDI Rendimento oO ur 176 352 528 704 880 1056 1232 1408 1584 1760 1936 2110 st oc w N 2632 2808 2951 3167 3343 3519 3655 4047 4223 4359 1000 Power W 500 Pot ncia Entrada Carregador 1 W Pot ncia Saida Carregador 1 W Pot ncia Entrada Carregador 2 W Pot ncia Saida Carregador 2 W Time s Fig 5 Current THD blue and Charger Efficiency red The inc
14. isplays for the registry of the parameters evolution in time and numeric indicators for the instantaneous values The battery and chargers tests were based on cycles of charge followed by a discharge The charge current starts at 30 A lowering by steps with the SOC increase The discharge current was different depending on the test 15 A 30 A and 60 A However it should be noted that the current 194 was not really constant since the discharge resistance was constant and the voltage decreases with the discharge Due the laboratory limitations the maximum continuous discharge current was 60 A Edit Operate Tools Browse Window Hel EO Tempo Entre Amostragens Ms STOP J 1000 azul 1 9 37 14 585 9 50 00 000 10 00 00 000 10 11 20 22 09 2010 09 2010 22 09 2010 22 09 20 Fig 2 LabVIEW Monitoring Program Visual Interface during a charge A charge or a discharge ends when at least one cell reaches the limit voltage values defined in the CellLog alarm The same limits were set for all Cel Log 8S devices in all tests 4 2 V for maximum voltage 2 7 V for minimum voltage and AV lt 0 5 V between cells Following are shown the experimental tests results for one charge As showed in Fig 3 the charge current is constant approximately 30A until the change of stage by the char
15. led a complete charge and discharge of the battery pack was done so that the cells are as balanced as possible The balancing system works according the user settings For these tests the settings were Cut off voltage 2 60V Low voltage 2 75V Nominal voltage 3 60V Charge voltage 4 00V Three cycles of charge and discharge were performed In all cycles the results for the cells voltages evolution stored energy and efficiency were identical to those obtained in the tests without the RT BMS system contrary to what was initially expected After analyzing the situation it was concluded that with a 30 A or 22 A charge current at PRZEGL D ELEKTROTECHNICZNY Electrical Review ISSN 0033 2097 R 88 NR 1a 2012 least with the used heatsinks the CBUs temperature rose significantly Fig 16 and they were not able to divert enough current to allow a much better cell charge balancing than with CellLog 8S l e when some cell s voltage starts to skyrocket the overall pack charge will soon finish as that cell will reach the voltage limit before the others are fully charged and so the system shuts the charger down Fig 16 Thermographic image of 3 CBUs during cell balancing Conclusions A LiFePQ battery pack was assembled and tested From this study the results led to the following main conclusions concerning 1 The cells and the CellLogs based interrupt system During the charging without a well adjusted RT
16. rease of the THDi is due to that mostly the high frequency harmonics assume a bigger influence when compared with the fundamental frequency wave in the lower load level The efficiency decrease on the other hand is due to the increase of the harmonics influence The charger global efficiency depends on the difference of the energy absorbed by the charger and the energy absorbed by the batteries Fig 6 2500 Abs Carregador Wh mE Abs Baterias W h 2000 TG 1500 Q 2 1000 Pl 2 500 Ww 0 Hn AEON ONOOA ONON ANIN mint Winn AV UEVE TATED PUTER TE O CON OTPFTOUON DTFONUWOO TNO N DxAAMmeMODAHMN AHMAD MN ONONNowommonoevvuonmrzZznwnaennmoe0QodrewernrtwA a nm ty no wy mor Wont nMnonr ODwwAaNnwtvooowwrdraannvouwwnoeoqoNnm Wh m SO NN NNN NOON OH TNH MM SF t t Time s Fig 6 Energy Absorbed by the Charger blue and Energy Absorbed by the Batteries red The chargers global results can be summarized as THDi min 95 THDi max 110 Instantaneous Efficiency 85 Global Efficiency 1820 4 2161 2 0 842 gt 84 2 Characterization of LiFePO4 Battery Pack For each test of charge or discharge the energy absorbed or supplied by the batteries was calculated by the developed system from the measured currents and voltages as well as the efficiency for each discharge current Fig 7 shows an example of the pack 1 first 16 cells charge voltage and the chargers output curr
17. sues and related costs are still being the main obstacles to overcome but significant progresses are being done in this area namely with the advent of lithium based batteries In order to supply a small but high performance EV and to deepen the knowledge 1 4 on the subject of charging systems for electric vehicles the authors at the Polytechnic Institute of Coimbra Engineering Institute of Coimbra IPC ISEC have set up a Lithium Iron Phosphate LiFePO4 battery bank and developed a study around it to make its characterization In this paper the technical specifications of the cells battery pack and chargers are firstly presented Then the set up of the batteries for utilization including the first charge and discharge cycle with over voltage and deep discharge under voltage protection is described It follows the description of the system and tests for battery pack and chargers characterization including the calculation of the instantaneous power the energies involved and performing energy quality analysis Finally a real time battery management system installation is presented along with some conclusions Specifications of cells battery pack and chargers The study develops around the TS LFP9OAHA 5 cells from Thunder Sky recently Winston Battery Limited with LiFePO4 cathode 3 0 3 3 V nominal voltage and capacity C 90Ah Fig 1 Standard charge discharge currents are 27 A 0 3 C maximum 3C 270 A maximum charge volt
18. to carefully check the individual cells behavior and balance the cells reducing their voltage and charge disequilibrium and put them in same initial condition For the pack first cells the first charge was made individually cell by cell using laboratory power supply sources witha 3A charge current However this was very time consuming and to accelerate the process it was done by groups of 4 cells or even 8 cells During these tasks precautions were taken to prevent the damage of the cells That was ensured through a monitoring and interruption system based on commercial electronic devices the CellLog 8S 6 cf Fig 15 Using only nine wires each one of these devices can measure show and store up to 8 cells voltages and transmit it to a computer using an USB connection This device possesses a system of adjustable alarms for maximum and minimum voltages levels a time alarm and an alarm for voltage difference between cells When an alarm situation is detected a solid state relay contact changes its state Making use of the alarm feature of the CellLog 8S devices an interruption circuit was built using relays push buttons and the CellLog alarm connector This system opens the power contactor when one alarm of any of the CellLogs triggers interrupting the charge or discharge process PRZEGL D ELEKTROTECHNICZNY Electrical Review ISSN 0033 2097 R 88 NR 1la 2012 193 However special care as to be taken as the CellLog alarm soli
Download Pdf Manuals
Related Search