SPEEDAM 2006 International Symposium on Power Electronics, Electrical Drives, Automation and Motion Energy Management Strategy for a parallel Hybrid Energy Management Strategy a parallel Electric Vehicle using for Fuzzy Logic Hybrid using Fuzzy Logic H.Electric Hannoun, D.Vehicle Diallo, Senior member, IEEE, C Marchand Laboratoire de Génie Electrique de Paris (LGEP) CNRS UMR 8507 Reference : Road Electrical Phone : +33169851664 – Fax : Topic +33169418318 – E-mailVehicles : hala.hannoun@lgep.supelec.fr Abstract- In this paper a fuzzy logic, rule based control strategy is proposed for a parallel hybrid electric vehicle (HEV) The strategy controls the amount of energy flow among components in order to satisfy the driver demand, optimize energy consumption and reduce polluting emissions After modelling powertrain components, the strategy is tested by simulation with three speed cycles An improvement of the ICE efficiency, therefore a reduction of the polluting gas, is obtained and revealed by the operating points of the ICE in the torque speed frame Keywords- Parallel hybrid electric vehicle, energy management strategy, fuzzy logic, optimization computing It presents the advantage to combine nonlinear control theories and linguistic knowledge The basic idea of a Fuzzy Logic Controller (FLC) is to formulate human intelligence and reasoning, which can be represented as a collection of if-then rules, in a way tractable for computers [4] FLC is tolerant to imprecise measurements and to component variability Previous research indicated that it’s suitable for hybrid vehicle control [5] This paper is built up as follows The vehicle topology will be described and analyzed in Section II The optimization strategy is presented in Section III and its performance will be evaluated by simulations in Section IV Conclusions are given in Section V II SYSTEM MODEL DESCRIPTION I INTRODUCTION The number of automobiles being introduced on the road is increasing every year so it’s adding to the pollution problem: Internal Combustion Engine (ICE) Vehicles are to blame for the major source of urban pollutions that cause the greenhouse effects leading to global warming To the environmental problem is also added an economic one inherent in the dependence on oil that may soon lead to a global crisis when the oil reserves in the world wane [1] Searching for new technologies that make vehicles less polluting and more economic have become a necessity: the electric vehicles appeared, initially, like a promising solution Their short driving distance and long recharging time for batteries combined with a high cost did not enable them to be competitive with respect to the conventional vehicle Another solution seems today to appear: hybrid vehicles The presence of the electric motor (EM) introduced additional degrees of freedom into the control technique of the propulsion system An effective management of the energy exchanges between various components (motors, batteries, etc) allows significant reduction of fuel consumption and pollutants emissions Considerable effort has been made to develop control strategies [2] The main objective is to increase the ICE efficiency From an analytical point of view, this means that a cost function has to be found including all the constraints such as battery state of charge and the EM ratings for example So the optimization methods, such as dynamic programming [3] are potential candidates but their implementation remains a tedious task and convergence is not always guaranteed Therefore the use of artificial intelligence techniques helps to avoid those traps In fact, fuzzy logic rules are easy to implement and are less time 1-4244-0194-1/06/$20.00 ©2006 IEEE S34 - 36 The proposed study has been done on the parallel HEV configuration In fact differing from the series hybrid, the parallel HEV allows both the (ICE) and the electric motor to deliver power in parallel to drive the wheels Since both the ICE and the electric motor are coupled to the drive shaft , the propulsion power may be supplied by the ICE alone, by the electric motor or by both (power assist) [6] Fig Block Diagram of a simple axis Parallel HEV We have chosen the simplest architecture in which the two motors are mounted on the same shaft and as a result they have exactly the same speed (fig 1) In following sections the models of each component of the powertrain, which will be used in our study, are briefly explained A The dynamic model of the vehicle The selected vehicle is of the segment M1 type (307, Xsara, Civic…) Its parameters are listed in Table I The vehicle is considered like a moving mass subjected to the tractive force FTR provided by the propulsion unit and to various efforts related to the road load FRL, such as FRL = fro + fao + fg [7] fro = f.m.g represents the rolling resistance of the tires and g is the gravitational acceleration constant fao = a v2 represents the aerodynamic drag force where a = 0.5ȡCxS is a constant and v is the vehicle speed (ȡ is the air density) fg = m.g.sinĮ represents the gravitational force where Į is the grade angle with respect to the horizon The vehicle speed is evaluated using the equation of dv motion, given by: FTR – FRL = m dt - B The internal combustion engine (ICE) Fig EM torque speed curve The ICE must be able to hold cruising speed because the EM is used for the starting phases and comes to assist the ICE when the requested power increases, i.e during acceleration For a 140 km/h cruising speed, a 55 kW engine suits Figure shows its efficiency map in the speed-torque frame and its maximum speed (475 rd/s) [8] The best efficiency curve defines the optimal engine operating points that implies the minimum fuel consumption; the desired situation is to operate the engine along that curve The electric motor is also designed to operate in generator mode in order to recover energy (regenerative braking for example) and charge the battery The characteristics of both modes are represented in fig a) Torque speed curves Fig Engine efficiency map [8] C The electric motor(EM) Acceleration from to 50 km/h in seconds is considered for sizing the EM A 37 kW induction motor has been adopted because beside the PMSM, it’s a good candidate to propel a HEV [6] The EM is represented as a state space model in the stator (Į, ȕ) frame As the speed is imposed by the shaft, a torque vector control strategy is applied The motor characteristic (fig 3) shows that the speed range is limited to 1175 rpm To assist the ICE, EM speed range needs to be extended up to 4500 rpm (475 rd/s) Therefore a flux weakening is introduced as it can be seen in fig b) Power speed curves Fig EM characteristics in motor and generator modes Fig b shows that the EM power limits are 30 kW in motor mode and 37 kW in generator mode D The battery There are many types of batteries and many factors that affect battery performance To predict the performance of batteries, many different mathematical models exist [9] S34 - 37 None of those models are completely accurate and include all necessary performance-effecting factors A Ni-MH battery type has been chosen; its characteristics are listed in Table II Its size, i.e the number of series (Ns) and parallel (Np) cells, is related to the induction motor data and the total vehicle weight The parameters of the battery electrical model are adjustable according to the type and the application Laboratory tests representative of the application concerned must be done in order to determine the accurate parameters Our model is limited to an internal electromotive force “E” in series with a resistance “R”, both being functions of the state of charge (SOC) The discharge characteristic (fig 5) is used to determine R (SOC) and E (SOC) Two polynomial functions are then calculated so that for every SOC, E and R can be computed (see fig 6) SOC is estimated according to the following equation: ³ Idt SOC( t 0) ; C * 3600 where SOC(t=0) is the initial state of charge and C=6.5 Ah SOC( t ) The open circuit voltage Voc can be evaluated according to following equation: I Voc Ns(E R ) , Where I is the current flowing in or Np out the battery For a better use and a longer lifetime of the battery, the optimization strategy must maintain a state of charge between 0.7 and 095 b) R (SOC) Fig Battery model characteristics E The gear box The gear box located between the primary (or motor) shaft and the secondary (or wheel) shaft is of a discrete type and is modeled by a ratio k that transforms the torque and speed Losses are not included in the model The number of ratio is determined by the relation between the engine and the vehicle speeds and their values depend on the ICE operation A five speed box is implemented for our application III FUZZY LOGIC CONTROLLER The optimization strategy objectives can be summarized as follows: - satisfy the conductor demand, - reduce the fuel consumption and consequently the CO2 emissions, - maintain the battery state of charge between 0.7 and 0.95 Therefore, the controller (fig.7) has three inputs (the requested power Pr calculated through the requested speed, SOC and vehicle speed V and two outputs the power PICE requested from the ICE and the gear ratio k The desired electric power PE is therefore : PE = Pr – PICE Fig Battery discharge characteristic Fig The fuzzy logic controller The inputs are defined by three membership functions respectively (μi for i=1 to 3), represented in fig 8, {L (low), N (normal), LM (low medium), M (medium), MH (medium high), H (High), Neg (Negative)} The controller a) E (SOC) S34 - 38 is of Sugeno type and therefore outputs are regular values and have no associated membership functions The ICE is the main power source for the propulsion system and has the lowest efficiency compared with the EM, it is reasonable to focus on optimizing engine efficiency [5] The ICE map used in this study (fig 2) shows that: - For a given speed, the efficiency increases as the PICE is higher - Efficiency is higher for speeds between 230 and 320 rd/s, corresponding to an ICE power between 30 and 50 kW always the ICE operating points on the optimal curve For example: 1- If motor speed = 250 rd/s, Pr = 20 kW and SOC < 0.95, the ICE operating point can be on the optimal curve PICE = 35 kW and PE = -15 kW(fig b) 2- On the other hand if motor speed = 450 rd/s, Pr = 20 kW and SOC < 0.95, the ICE cannot be optimized with the best efficiency (PICE = 55 kW), because at this speed the EM can absorb a maximum of 30 kW Another important criteria is to avoid engine operation in low power areas where the efficiency is very low (as indicated in Fig 2) A threshold is set up so the engine turns on only when power request exceeds 10 kW In that case, the vehicle is propelled only by the electric motor A base of thirteen rules is established One of the rules is for example: “If Pr is M and SOC is L then PICE is 55” i.e the ICE supplies its maximum power when the requested power is medium and the state of charge is low Another rule is given by: “If Pr is Neg then PICE is 0” i.e during the braking phase, the ICE is shut down, only the EM is running in order to recover power IV SIMULATION RESULTS The strategy is validated by simulation using MatlabSimulink software package Three different speed cycles are implemented in order to illustrate the different operating modes: - The ICE provides alone the totality of the requested power - The EM provides alone the totality of the requested power - The power assist mode - The batteries are recharged through the additional power supplied by the ICE - Take advantage of the braking to recharge the batteries Fig Membership Functions of the FLC The main goal is then to set the ICE operation in its best efficiency region This improves the overall efficiency of the powertrain The controller must always try to operate the engine along the optimal curve or as close as possible This operation does not necessarily meet the diver’s demand so the battery becomes an extra load to adjust the engine’s operation so that it’s close to the best efficiency line In this case, the battery is charged through the EM acting as a generator But it is also necessary to respect important constraints on the battery and on the electric power limitations of the motor in generator mode So it will be difficult to have S34 - 39 The first use case shown in figure is extracted from the European Normalised Cycle (93/116) Figure 10 shows the requested power and the ICE power Below 10kW, only the EM propels the vehicle and the ICE is idle It’s turned on as soon as Pr exceeds 10 kW The FLC generates the power requested from the ICE and the comparison of the operating points without and with hybridizing is illustrated in fig 11 It shows that the ICE operating points lies in a higher efficiency area The power surplus is used to recharge the battery through the EM if SOC is below 0.95 and to supply the auxiliaries on board the vehicle If SOC >0.95 and the available power is superior to the need of the auxiliaries, we can see in fig 11b that the operating points of the ICE are no longer on the optimal curve Nevertheless the efficiency is still better than in a conventional vehicle b) Operating points of ICE with hybridizing Fig 11 Operating points of the ICE The last cycle illustrates the bi – propulsion or power assist mode when the two motors contribute to propel the vehicle Between 41.5 and 45 seconds, the requested power exceeds the ICE capacity (fig 13) therefore impelling the EM to furnish the rest Figure 13 shows the requested power and how the two motors operate Different operating modes can also be noticed The notation “+EM” corresponds to the EM supplying power whereas “–EM” corresponds to the EM running as a generator Fig Reference and actual vehicle speeds Fig 10 Power curves Fig 12 Reference and actual vehicle speeds a) Operating points of ICE alone S34 - 40 TABLE I Vehicle parameters m : total weight f : coefficient of rolling resistance S : frontal area Cx : Aerodynamic coefficient wheel radius 1300 Kg 0,0133 2,61 m2 0,32 0,32 m TABLE II NI-MH battery specifications Nominal Voltage Nominal Capacity Specific Power Specific Energy Weight Fig.13 Power curves Dimensions 7,2 V 6,5 Ah 1300W/kg 46Wh/kg 1040g 19,6(W) X 106(H) X 285(L) REFERENCES Fig 14 ICE operating points Again figure 14 shows that the strategy allows to operate the ICE in a higher efficiency area than the conventional vehicle V CONCLUSION This paper has presented a fuzzy logic and rule based energy management strategy for a parallel HEV The efficiency map of the internal combustion engine has been used to design the controller According to the requested power, the vehicle speed and the battery state of charge, the fuzzy controller chooses the best power split between the ICE and the EM The strategy has been validated by simulation A comparison of the ICE operating points with those of the conventional vehicle shows how the hybridizing added to a suitable management strategy can improve considerably the efficiency and thus reduce pollutant emissions The strategy has the potential to be implemented in a real vehicle control system S34 - 41 [1] Husain, Electric and Hybrid Vehicles: Design Fundamentals, CRC PRESS, 2003 [2] B K Powell, K E Bailey, and S R Cikanek, “Dynamic modelling and control of hybrid electric vehicle powertrain systems”, IEEE Contr Syst Mag., pp 17-33, Oct 1998 [3] A Sciaretta, M Back, L Guzella, “ Optimal Control of Parallel Hybrid Electric Vehicles, “ IEEE Trans Contr Syst Technol., vol 12, n° 3, pp 352-363, May 2004 [4] N Schouten, M Salman, and N Kheir, “Fuzzy logic control for parallel hybrid vehicules, ”IEEE Trans Contr Syst.Technol., vol 10, pp 460-468, May 2002 [5] X He, M Parten, T Maxwell, “Energy Management Strategies for a Hybrid Electric Vehicle, ” IEEE Vehicular power electronics and propulsion conference, CD ROM, pp 536-540, September 2005, Chicago, IL [6] M Zeroualia, M E H Benbouzid and D Diallo, “Electric Motor Drive Selection Issues for HEV Propulsion System : a comparative Study ”, IEEE Vehicular power and propulsion conference, CD ROM, pp 280-287, September 2005, Chicago, IL [7] M Ehsany, K M Rahman, and H A Toliyat, “Propulsion system design of electric and hybrid vehicule application”, IEEE Trans Ind Electron., vol 44, pp 19-27 Feb 1997 [8] Naim A Kheir, Mutasim A Salman, Niels J Schouten “Emissions and fuel economy trade-off for hybrid vehicles using fuzzy logic”, Mathematics and Computers in simulation 66 (2004), pp 155-172 [9] P.H Mellor, N Schofield, A.J Brown, D.Howe, “Assessment of supercapacitor/flywheel and battery EV traction system”, ISATA 2000, pp 235-242, session ‘Electric/Powertrain’, Dublin, September 2000 ... conventional vehicle shows how the hybridizing added to a suitable management strategy can improve considerably the efficiency and thus reduce pollutant emissions The strategy has the potential to... operating points Again figure 14 shows that the strategy allows to operate the ICE in a higher efficiency area than the conventional vehicle V CONCLUSION This paper has presented a fuzzy logic and... battery For a better use and a longer lifetime of the battery, the optimization strategy must maintain a state of charge between 0.7 and 095 b) R (SOC) Fig Battery model characteristics E The gear