model-based development of future small evs … · model-based development of future small evs...

Model-based Development of Future Small EVs using Modelica

Yutaka Hirano Shintaro Inoue Junya OtaToyota Motor Corporation, Future Project Division

1200 Mishuku, Susono, Shizuoka, 410-1193 JAPAN{yutaka_hirano, shintaro_inoue_aa, junya_ota}@mail.toyota.co.jp

Abstract

To cope with demands for future low carbon society,development of new-type small electric vehicles(EVs) becomes very active. To reduce the energyconsumption in various actual driving conditions,considering overall running resistance such as aero-dynamic resistance, tire rolling resistance includingcornering drag, mechanical and electrical losses, etc.will be necessary. On the other hand, to cope withreduced stability against external disturbances such asside wind because of the light weight, it was clarifiedthat additional control of direct yaw moment is effec-tive. In this paper, model-based development of anew electric vehicle using Modelica is described. Fullvehicle model considering both vehicle dynamics andenergy consumption was developed and utilized toinvestigate the best possible solutions for both basicdesign of the vehicle and design of the control system.

Keywords: Future electric vehicles; Stability andHandling Performance; Energy Consumption

1 Introduction

To cope with future mobility society, development ofmany new concept vehicles is increasingly active inrecent years. Figure 1 shows a new EU regulationabout light weight vehicles [1]. Those vehicles havecharacteristics of smaller size, lighter weight, lessnumber of passengers than conventional vehicles.Also those vehicles tend to be equipped with lowerRRC (Rolling Resistance Coefficients) tires and newdriving systems mainly using electric motors toachieve less emission and less energy consumption.On the other hand, Toyota has a vision about futureeco-cars as shown in Figure 2. Toyota thinks EVs are

suitable as short-distance mobility though there is apossibility of extending the driving range using rangeextender devices such as small combustion engine,additional battery and so on. In this paper, model-based-development of a new vehicle using Modelicais described. The models were developed based onVehicle Dynamics Library (VDL) of Dymola.

Figure 1: New EU Regulation "Light-category vehi-cles" [1]

Figure 2: Toyota’s scenario about future eco-cars

DOI10.3384/ECP1409663

Proceedings of the 10th International ModelicaConferenceMarch 10-12, 2014, Lund, Sweden

63

2 Modeling and simulation studies ofthe future vehicle

2.1 Target vehicle

Table 1: Specifications of the target vehicle

Target plan L7Be

Vehicle weight < 600 kg < 400* kg

Passengers 4 2

Max. Payloadincl. passengers

300 kg 200 kg

Rated power 25 kW < 15 kW

Max. speed 120 km/h > 45 km/h

Driving range > 100 km -

(* weight without batteries)

Table 1 shows comparison of specifications betweenour plan and EU regulation L7Be. Our aim is to de-velop a heavier and more powerful vehicle with morepassengers than L7Be considering actual usefulness.

2.2 Simulation studies about basic specifications

To consider energy consumption, handling, stability,ride comfort and NVH (noise, vibration, harshness)performances of holistic vehicle, a full-vehicle modelincluding mechanics, electronics, vehicle dynamicsand control was made using Dymola. Moreover, amodel of a new drive train system such as torque vec-toring differential gear was developed and connectedinto the full-vehicle model. Figure 3 shows an exam-ple of the full-vehicle model. Details of the modelwill be explained later.

Top layer

ControllerControllerTorque vectoringTorque vectoringdifferential geardifferential gear

Suspension and tiresSuspension and tires

Figure 3: An example of full-vehicle model

Power consumption of each system was calculatedsimultaneously and was used for the investigation ofgood balance of energy consumption and vehicle per-formances. At first, total power of resistances actingon the vehicle was calculated by following equations[2].

Total resistance power:

sxsyarrrv PPPPP (1)

Rolling resistance power:

VmgP rrr (2)

Aerodynamic resistance power:

VVACP Dar 2/2 (3)

Cornering resistance power:

VgmAC

d

C

dP y

pr

r

pf

f

sy

/2

VgmAC

y

p

/

1 2 (4)

Longitudinal resistance power:

VmgmAP xsx )sin( (5)

Here

r : rolling resistance coefficient (RRC) ,

g: acceleration of gravity [m/s2],,

m: vehicle mass [kg],

V: vehicle speed [m/s],

: air density [kg/m3],

A: vehicle frontal area [m2],

DC : aerodynamic resistance coefficient,

fd : front weight distribution ratio,

rd : rear weight distribution ratio,

pfC : front normalized cornering power [1/rad],

prC : rear normalized cornering power [1/rad],

pC : average normalized cornering power [1/rad],

yA : lateral acceleration [m/s2],

xA : longitudinal acceleration [m/s2],

: road inclination [rad].


64 Proceedings of the 10th International ModelicaConferenceMarch 10-12, 2014, Lund, Sweden

DOI10.3384/ECP1409663

0

0.5

1

1.5

2

2.5

3

Res

ista

nce

pow

er[k

W]

PsxPsyParPrr

Straight Cornering (0.4G)

0

0.5

1

1.5

2

2.5

3

Res

ista

nce

pow

er[k

W]

PsxPsyParPrr

Straight Cornering (0.4G)

Figure 4: Comparison of resistance powers whiledriving straight and cornering (V = 30[km/h]) bysimulation

0

1

2

3

4

5

6

Y4

5[m

]

Conventionalvehicle

New smallvehicle

(CP=15) (CP=15)

Target

(CP=24) (CP=12)

Figure 5: Simulation results of parameter study forstability against side wind between conventional ve-hicle and the new small vehicle

Figure 4 shows an example of a comparison of resis-tance powers between straight driving and cornering.It became clear that rolling resistance and corneringresistance were rather big and reducing those resis-tances was essential to reduce the power consumption.

From equations (2), (4) and (5), it is understood thatdecreasing vehicle mass and tire RRC and also in-creasing tire cornering power (CP, normalized by tirecontact load) are effective to reduce the total resis-tance power and improve the energy consumption ofthe vehicle. However, in general, decreasing RRCtends to result in decrease of CP for ordinary tires.Moreover, it is expected that decreasing vehicle mass

will result in reduced vehicle stability against externaldisturbances such as side wind. Figure 5 shows a re-sult of parameter study for evaluating stability againstside wind by Dymola. Evaluation criteria (Y45[m]) inFigure 5 is the lateral deviation while driving at 120km/h and pass a zone of 45m length with the sidewind of 20m/s. In Figure 5, comparison between con-ventional vehicle (m = 1050[kg]) and the new small

vehicle (m = 600[kg]) are shown. Also for the newsmall vehicle, some levels of normalized CP wereresearched. It became clear that the light-weight smallvehicle is affected more than the conventional vehicleby side wind, and sensitivity of normalized CP valueagainst the disturbance is very small. From aboveinvestigations it is indicated that developing new tirewhich can realize both low RRC and high CP value isnecessary for reducing energy consumption. Also forcoping with improving vehicle stability against exter-nal disturbances for such small vehicles, additionalcontrol of vehicle dynamics such as direct yaw mo-ment control is considered to be necessary.

3 Development of necessary items toimprove holistic performance of thenew small vehicle

3.1 New suspension system using tires with lowRRC and high CP value

As mentioned in the previous section, the develop-ment of new tires for which both low RRC and highCP value can be realized will be necessary for reduc-ing overall energy consumption. For this purpose, anew concept of tires called Large and Narrow (L&N)Concept was developed by Bridgestone [3]. It hascharacteristics of larger overall diameter, narrowersection width and higher inflation air pressure thanconventional tires. Figure 6 (cited from [3]) showscomparison of RRC and CP measurement data be-tween ordinary reference tire, L&N tire and L&W(Large and Wide) tire representing current high per-formance tire. It can be seen that L&N tire has goodbalance of lower RRC and higher CP at high inflationpressure of 320 kPa. Thus, it was decided to adoptL&N tires for our new vehicle. Figure 7 shows acomparison of overall cornering resistance force (Cr)calculated by the equation (6) when running on a con-stant radius corner at lateral acceleration of 0.4G.

Session 1B: Automotive Applications 1

DOI10.3384/ECP1409663


65

Figure 6: Measurement data of RRC and CP atLoad=3.5kN [3]

m:600kgRRC:6CP:16.1

m:600kgRRC:6CP:13

Co

rner

ing

Res

ista

nce

[N]

43%down

0

50

100

150

200

250

17%down3%down

m:1050KgRRC:9.5CP:16.1

Light weight

Conventionalvehicle New vehicle

LowRRC tire

L&N tire

m:600KgRRC:9.5CP:16.1

Figure 7: Comparison of cornering resistance force bysimulation

py CgmACr /2 [N] (6)

Comparison is done between conventional vehicleand new small vehicle with ordinary tire, low RRC(but low CP) tire and L&N tire. The effect of lowweight and L&N tire to reduce the cornering resis-tance and thus energy consumption was proved by thesimulation.However, there still are remaining problems for ap-plying L&N tires for the new vehicle. Because of lar-ger overall size, it has larger rotating inertia resultingin larger drive-train vibration and less controllabilityof driving torque than conventional vehicle. Alsohigher inflation pressure results in higher verticalstiffness and it is suggested to affect ride comfort andNVH (Noise, Vibration and Harshness). Thus, newsuspension design is supposed to be necessary. Tocope with those problems are future works.

3.2 Active yaw moment control by torque vec-toring system

As mentioned above, it became clear that active yawmoment control to cope with external disturbanceswas indispensable for small light-weight vehicles. Toresearch the best solution of this function, bench-marking of existing torque vectoring systems wereperformed using simulation by Dymola at first. Con-sidering the space for mounting and also controllabil-ity, TUM:MUTE type system[4] was investigatedfurther. In this system, a main motor connected toouter ring gear of the differential planetary gear setcontrols total driving torque. On the other hand, acontrol motor connected with an input shaft of controlgear sets controls torque distribution of left and right



DOI10.3384/ECP1409663

wheel. The function of torque distribution was con-firmed by Dymola simulation as shown in Figure 8. Itbecame clear that torque distribution ratio can bechanged from 50:50 to both of 0:100 and 100:0 andmore by increasing the input torque of the controlmotor. By this simulation, also energy consumptionof both main motor and control motor was able tocalculate as well as mechanical transient motion.Finally an example of desired yaw rate feedback con-trol for the torque distribution ratio was tested. Mainmotor torque (Tm) was decided by PI feedback con-trol of difference between desired value and actualvalue of the vehicle speed by following equation.

dtVVKVVKT refiVrefpVm )()( (7)

where

KpV: Proportional feedback gainKiV: Integral feedback gainVref: Desired vehicle speedV: Actual vehicle speed

On the other, control motor torque (Tc) was calcu-lated by PI feedback control of difference betweendesired yaw rate and actual yaw rate as followingequation.

dtrrKrrKT refirrefprc )()( (8)

-400

-300

-200

-100

0

100

200

300

400

0 2 4 6

Control motor power [kW]

To

rqu

ed

istr

ibu

tio

nra

tio

[%]

Right

Left

Main motor

+150%

-150%

Main motor

+150%

-150%

+300%

-300%

Figure 8: Simulation result of torque distribution ratioVS control motor power for torque vectoring differ-ential gear system of MUTE type

where

Kpr: Proportional feedback gainKir: Integral feedback gainrref: Desired yaw rater: Actual yaw rate

and the desired yaw rate was calculated as below.

s

s

sref

sT

Kr

1(9)

rfrffrr

rfrf

sccaaamVca

VccaaK

)(

)(2

(10)

frf

f

scaa

VmaT

)( (11)

Here,

δs : Steering input angle at front tireaf : Longitudinal distance between front wheel andCG (Centre of gravity)ar : Longitudinal distance between rear wheel andCGcf : Cornering stiffness of front two tyrescr : Cornering stiffness of rear two tyres.

Figure 9 shows a simulation result of the side windtest. Lateral deviation against the side wind (Y45) andenergy consumption of main motor and control motorwere calculated for the cases of no control, only Pfeedback control and PI feedback control. The con-tradiction between vehicle stability and energy con-sumption of the control motor was confirmed. Finally,development of a new torque vectoring differentialgear based on parallel planetary gear sets was decided.Numerical consideration as above will enable us todesign the best solution for the practical design of thesystems both in mechanical aspect and electrical as-pect.


DOI10.3384/ECP1409663


67

No Control P_Control

Target

0

0.5

1

1.5

2

2.5

3

3.5

4

Y45

[m]

PI_ControlNo Control P_Control

Target

0

0.5

1

1.5

2

2.5

3

3.5

4

Y45

[m]

PI_Control

<Lateral deviation against side wind [m]>

<Main Motor Power [kW]>

<Control Motor Power [kW]>

Side wind

Side wind

Figure 9: Simulation results of side wind test betweenNo control, P feedback control and PI feedback con-trol of torque vectoring differential gear

3.3 Electric regeneration system of brakingforce

0

1

2

3

4

5

6

100 150 200 250 300Time (s)

Po

wer

(kW

)

Rolling resistance [kW]Cornering resistance [kW]Aerodynamic resistance [kW]

0 50

Figure 10: An example of calculation of each resis-tance power when driving on a winding circuit road

-150

-100

-50

0

50

100

150

200

250

DriveBreak DriveBreak DriveBreak DriveBreak

JC08 US06 Defensive Normal

Ener

gy(W

h/km

)

Winding circuit courseFuel economy mode

-150

-100

-50

0

50

100

150

200

250

DriveBreak DriveBreak DriveBreak DriveBreak

JC08 US06 Defensive Normal

Ener

gy(W

h/km

)

Winding circuit courseFuel economy mode

Acceleration andRoad inclinationRolling resistanceCornering resistanceAerodynamic resistance

Mechanical brakingRegenerative breaking

Possibility of recuperatingenergy by regenerativebreaking

Acceleration andRoad inclinationRolling resistanceCornering resistanceAerodynamic resistance

Mechanical brakingRegenerative breaking

Possibility of recuperatingenergy by regenerativebreaking

Figure 11: Comparison of driving and breaking en-ergy for some examples of driving modes

Utilizing electric regeneration of braking force bymotor is effective to improve the energy consumption.In the planning phase of a new vehicle, it is necessaryto decide proper size of battery capacity for the re-generation. For this purpose, realistic simulation ofactual driving scenes is necessary. IPG CarMaker wasused to calculate the vehicle speed, longitudinal andlateral acceleration, road decline and corner radius foractual roads. Using above data, necessary drivingpower and overall resistance power by the equation



DOI10.3384/ECP1409663

(1) were calculated. Figure 10 shows an example ofresistance power of each resistance force while driv-ing on a winding circuit road. Using these results,overall driving energy and braking energy for someexamples of driving mode were calculated as shownin Figure 11. Here, JC08 and US06 are regulation ofdriving modes for measuring fuel consumption inJapan and USA respectively. Figure 12 shows distri-bution of vehicle speed and longitudinal accelerationfor both modes. As seen in Figure 12, US06 modeuses higher vehicle speed and acceleration than JC08mode and results in more driving power and brakingpower appearing in Figure 11. Also in Figure 11,comparison of two driving styles for a winding circuitroad is shown. It is observed that the defensive driv-ing style needs less power than the normal drivingstyle. It is assumed that breaking power less than15kWcan be recuperated by the regeneration of motorin this example. It was understood that there is re-mained braking power which can be recuperated ifthe regeneration ability of battery system is largeenough in the case of circuit road driving. In this ex-ample of Figure 11, these braking powers are con-sumed by the mechanical breaks and wasted. Figure13 shows a result of simulation to calculate possibleelectricity consumption value in the cases of usingbattery systems whose regeneration ability are 15kW,40kW and 70kW respectively. It became possible toestimate how large battery capacity was necessary toimprove the energy consumption in each driving cas-es.

Consequently it was proved that these simulationswere very useful to decide the proper specificationsof a new vehicle in the planning phase.

4 Conclusions

For the investigation of overall vehicle specificationsand system structures, a holistic vehicle model in-cluding mechanics, electrics, electronics, vehicle dy-namics and control was made and utilized. It wasproved that such a holistic model was very useful toinvestigate the proper specifications and system con-structions in the phase of early stage of vehicle devel-opment. Modelica was very suited to make such amulti-discipline and multi-domain investigation bymodel-based development.

-0.4-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80 100 120 140

Vehicle speed (km/h)

Lo

ngit

udin

alac

cele

rati

on(G

)

JC08

US06

-0.4-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80 100 120 140

Vehicle speed (km/h)

Lo

ngit

udin

alac

cele

rati

on(G

)

JC08

US06

Figure 12: Vehicle speed VS longitudinal accelera-tion for JC08 and US06 modes

0

5

10

15

20

Defensive Normal

JC08 Winding circuit courseEle

ctri

city

cons

umpt

ion

(km

/kW

h)-15 kW-40kW-70kW

0

5

10

15

20

Defensive Normal

JC08 Winding circuit courseEle

ctri

city

cons

umpt

ion

(km

/kW

h)-15 kW-40kW-70kW

-15 kW-40kW-70kW

Figure 13: Comparison of estimated electricity con-sumption when changing the battery’s ability of re-generation

References

[1] Informal document GRB-55-15 (55th GRB,7-9 February 2012, agenda item 7(b)), Euro-pean Commission, Enterprise and Industry

[2] Kobayashi T., Katsuyama E., Sugiura G., OnoE., Yamamoto M.: “A research about drivingforce distribution control and energy con-sumption while cornering”, Proceeding of2013 JSAE Annual Congress (Spring), 352-20135393, 2013 (in Japanese)

[3] Kuwayama I., Matsumoto H., Heguri H.:“Development of a next-generation-size tirefor eco-friendly vehicles”, Proceeding of


DOI10.3384/ECP1409663


69

Chassis.tech plus | 4th International MunichChassis Symposium 2013, pp.623-644, 2013

[4] Höhn B., Stahl K., Wirth C., Kurth F., Lienk-amp M., Wiesbeck F.: “ElectromechanicalPower Train with Torque Vectoring for theElectric Vehicle MUTE of the TU München”,Getriebe in Fahrzeugen 2011 Effizien-zsteigerung im Antrieb, 7./8./ Juni (2011),Friedrichshafen. VDI-Berichte Nr. 2130, 2011



DOI10.3384/ECP1409663

model-based development of future small evs … · model-based development of future small evs...

Documents