a novel energy management strategy for series hybrid...

Research ArticleA Novel Energy Management Strategy for Series HybridElectric Rescue Vehicle

Pei Li Jun Yan Qunzhang Tu Ming Pan and Jinhong Xue

Field Engineering College Army Engineering University of PLA Nanjing 210007 China

Correspondence should be addressed to Qunzhang Tu tqzlhnj126com

Received 3 April 2018 Revised 26 July 2018 Accepted 27 August 2018 Published 29 October 2018

Academic Editor Francesco Braghin

Copyright copy 2018 Pei Li et al This is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

The performance and fuel consumption of hybrid electric vehicle heavily depend on the EMS (energy management strategy)Thispaper presents a novel EMS for a series hybrid electric rescue vehicle Firstly considering the working characteristics of engine andbattery the EMS combining logic threshold and fuzzy control is proposed Secondly a fuzzy control optimization method based onIQGA (improved quantum genetic algorithm) is designed to achieve better fuel efficiencyThen the modeling and simulation arecompleted by using MATLABSimulink the results demonstrate that the fuel consumption can be decreased by 517 after IQGAoptimization and that the optimization effect of IQGA is better than that of GA (genetic algorithm) and QGA (quantum geneticalgorithm) Finally the HILS (hardware in loop simulation) platform is constructedwith dSPACE the HILS experiment shows thatthe proposed EMS can effectively improve the vehicle working efficiency which can be applied to practical application

1 Introduction

In recent years natural disasters (such as earthquakes debrisflow and heavy snowfall) have often occurred causinginestimable loss to the peoplersquos life and property As animportant part of the emergency response system rescuevehicles have a direct impact on the efficiency of disaster reliefoperations The application of hybrid electric technologyto rescue vehicles has aroused peoplersquos attention by whichthe vehicle mobility and fuel efficiency can be significantlyimproved At the same time the generator can be used toprovide power for lighting equipment rescue equipmentand other electrical equipment so the vehicle adaptabilityin harsh environment can be enhanced However the EMSplays a key role in the performance of hybrid electric system[1] which can be roughly classified into the following threecategories

(1) Logic Threshold Control Strategy This strategy which iswidely used in many automobile enterprises is simple andeasy to develop mainly including the ldquothermostatrdquo controlstrategy and power following control strategy [2] By settingthe logic threshold to establish control rules each powercomponent can work in the high efficient area as much as

possible However the setting of threshold value dependson the experiences of experts and the optimal control effectcannot be guaranteed [3 4]

(2) Optimization Based Control Strategy By defining objectivefunction combining the constraint conditions the controlobjective can be optimized by mathematical algorithm Thisstrategy can be subdivided into global optimization controlstrategy and instantaneous optimization control strategyTheformer is designed based on given driving cycle whichis often used for vehicles with fixed routes such as busesor commuter cars Besides it is also used as a criterionto evaluate other control strategies Because the calculatingquantity is large and complex the optimization process can-not be carried out onlineThe representative strategiesmainlyinclude dynamic programming (DP) based strategy [5 6]Pontryaginrsquos minimum principle (PMP) based strategy [7 8]Quadratic programming (QP) based strategy [9 10] dividerectangle (DIRECT) algorithm based strategy [11ndash13] andconvex optimization based strategy [14] The latter is aimedat achieving optimal energy management in instantaneousstate which is not restricted by the specific driving cycle Thecomputation is relatively small and can be applied online

HindawiMathematical Problems in EngineeringVolume 2018 Article ID 8450213 15 pageshttpsdoiorg10115520188450213

2 Mathematical Problems in Engineering

Motorcontroller

General controller

DC bus

CAN busElectric linkMechanical link

Brakingcontroller

Motor

Brake

Brake

Brake

Brake

Finaldrive MotorIGPU

controller

ACDC

IGPU

BMS

Battery pack

DCDC

Motorcontroller

Finaldrive

Figure 1 Structure diagram of the hybrid rescue vehicle

such as fuel consumptionminimization strategy (ECMS) [15ndash17] and model predictive control strategy (MPC) [18ndash20]However the instantaneous optimization belongs to localoptimization and the optimization effect depends on theprecision of the mathematical model

(3) Fuzzy Logic Control Strategy With good real-time per-formance and robustness [21 22] the fuzzy logic controlstrategy does not rely on precisemathematical models whichis suitable for dealing with the complex nonlinear andtime-varying problems In practical applications the speedSOC required torque and power are often used as inputvariables [23ndash25] and the power ratio of the engine andthe battery is used as the output variable However theformulation of control logic is mainly based on engineeringexperience and it is difficult to ensure the optimal powerdistribution To achieve better performance many strategiescombining fuzzy control method with other control methodshave been proposed by some scholars In [26] a fuzzycontroller was used to adjust the equivalent factors of ECMSand the fuel economy of the hybrid electric heavy-dutyvehicle was improved In [27] the DP algorithm was usedto optimize the performance of ISG hybrid electric vehicleand combined with the optimization results a more practicalfuzzy controller was proposed In [28ndash30] some intelligentalgorithms such as genetic algorithm and particle swarmoptimization were used to optimize the membership func-tion or control rules of the fuzzy controller In [31ndash33] themixed-fuzzy control strategy was designed by using neuralnetwork working condition identificationmachine learningand other intelligent techniques However the control effectsof the above strategies still exist a certain improvement space

Considering the advantages and disadvantages of theabove three categories of methods a novel EMS is proposed

to improve the fuel economy of a series hybrid electricrescue vehicle in this paper First combined with thresholdcontrol and fuzzy control the EMS is initially designedaccording the working characteristics of engine and batteryThen an improved quantum genetic algorithm with betterconvergence performance is proposed and applied to optimalthe fuzzy controller The structure of the paper is as follows inSection 2 the configuration of a series hybrid electric rescuevehicle is introduced in Section 3 the EMS of the wholevehicle is studied in Section 4 the method for optimizingfuzzy control based on IQGA is designed in Section 5 theoffline simulation andHILS experiment are described finallythe paper is concluded in Section 6

2 Configuration of the Series HybridElectric Rescue Vehicle

In series hybrid electric system there is no mechanicalconnection between the engine and the driving wheels thespeed and torque of the engine can be controlled to anyoperating point on its speed-torque map So by optimizingcontrol the fuel consumption and emission of the enginecan be improved As shown in Figure 1 the four-wheel drivestructure with two motors is adopted to the rescue vehicle inthis paper which can not only increase the vehicle power per-formance but also provide better regenerative braking effectThe electric drive system mainly includes general controllerinternal-combustion-engine-generator power unit (IGPU)and its controller battery pack and its management system(BMS) motors and motor controllers rectifier and DCDCconverter The general controller is the core componentwhich is used to coordinate the power distribution of IGPUand battery pack according to the driver instruction vehicle

Mathematical Problems in Engineering 3

Table 1 Specific parameters of the hybrid rescue vehicle

Name ValueVehicle mass 14000kgFrontal area 625m2

Final drive ratio 79Wheel radius 072mGravity center height 1672mmDistance from gravity center to front wheel 3164mmWheel base 5200mmPeak power of IGPU 260KWPeak power of motor 150KWPeak torque of motor 2400NmBattery capacity 92kWh

running state and other feedback information As an auxiliaryenergy source the battery pack provides additional powerwhen the IGPU output power is insufficient such as thesituation of vehicle uphill and acceleration The two motorscan be used not only to provide driving force but also towork as a generator to charge the battery when the vehicleis braking The specific parameters of the vehicle are shownin Table 1

3 Energy Management Strategy Design

Under the condition of satisfying vehicle power performancehow to coordinate the energy distribution between IGPUand battery to realize the optimization of vehicle efficiencyis a primary problem that needed to be resolved for EMSThe following principles are taken into consideration inthe formulation of EMS (1) In order to improve engineefficiency the minimum fuel consumption curve controlmethod is adopted (2) In order to make the battery not onlyable to provide power in time but also able to recover theregenerative braking energy effectively the battery SOC iscontrolled in the high efficiency area and a certain margin isreserved for charging (3) On the premise of braking stabilityin order to make maximum use of the two motors to recoverbraking energy the braking force distribution method basedon I curve is adopted

Through the experiments the IGPU fuel consumptioncharacteristics and battery internal resistance characteristicsare obtained as shown in Figures 2 and 3 respectively It isknown that when the IGPU output power is less than 50KWthe fuel economy is bad and when SOC is in the range of05sim09 the battery working efficiency is high Therefore theIGPU output power is controlled within the range of 50sim260KW and the battery SOC is controlled within the range of05sim08 (SOClow=05 SOChigh=08) According to the drivingmode switch battery SOC and required power the vehicleworking state is divided into four operating modes pureelectric traction mode IGPU traction mode hybrid tractionmode and braking mode Each mode switches by the setthreshold and the flowchart of the EMS is shown in Figure 4

Engine speed (rmin)

Gen

erat

or p

ower

(KW

)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259266

273

273

280308

800 1000 1200 1400 1600 1800 2000

50

100

150

200

250

300

350

Fuel consumption rate (gkWh)Maximum output power curveOptimum operating line

Figure 2 MAP of the IGPU

ChargeDischarge

01 02 03 04 05 06 07 08 09 10SOC

001

002

003

004

005

006

007

008

009

01Re

sista

nce(

Ω)

Figure 3 Battery internal resistance characteristics

119875119903119890119902 is the required power of two motors 1198751015840119892 is thetarget power of IGPU 1198751015840119887 is the target power of batteryand 1198751015840119877 is the target power of brake resistance Define 119875119888maxis the maximum charging power of battery and 119875119889max isthe maximum discharge power of battery Assume that thecharging current and power of the battery are negative andthat the discharge current and power of the battery arepositive the energy distribution of different modes is asfollows

31 Pure Electric Traction Mode In this mode the batteryalone supplies its power to meet the power demand Andthis mode includes two cases (module 1 and module 2 inFigure 4) when the pure electric switch is opened the vehicleis in pure electric drive mode A when 0 le 119875119903119890119902 le 50KW andSOC gt SOClow the vehicle is in pure electric drive mode


Start

Pure electric

Electric mode A

Electric mode B

IGPU traction mode A

IGPU traction mode B

Hybrid mode A

Hybrid mode B

Braking mode A

Braking mode B

Return

Yes

No

Yes

Yes

No

No

NoYes

No

Yes

No

Yes

YesNo

Calculate Preq

SOC gt３ＦＩＱ

SOC gt３ＦＩＱ

Preq lt0

SOCge ３ＦＩＱ

SOCge ３ＢＣＡＢ

Output Pg P

b PR

Preqgt50KW

①

⑧

⑦

④ ⑥

⑤

②

③

Figure 4 Flowchart of the EMS

B Both of these two cases are satisfied with the followingformula

1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 0 (1)

32 IGPU Traction Mode In this mode IGPU not onlyprovides power to propel the vehicle but also provides powerto charge the batteryThere are also two cases included in thismode (module 3 and module 4 in Figure 4)When 0 le 119875119903119890119902 le50KW and SOC le SOClow the vehicle is in IGPU tractionmode A and the battery is charged with the maximumcharging current

1198751015840119887 = 119875119888max

1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (2)

When 119875119903119890119902 gt50KW and SOCle SOClow the vehicle is inIGPU traction mode B

1198751015840119887 = max (119875119888max 119875119903119890119902 minus 260)1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (3)

33 Hybrid Traction Mode This mode is also divided intotwo cases (module 5 and module 6 in Figure 4) When119875119903119890119902 gt50KW and SOCge SOChigh the vehicle is in hybridtraction mode A at this time the required power is mainlysupplied by the battery and the additional power is supple-mented by IGPU that is

1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (4)

When 119875119903119890119902 gt50KW and SOClow lt SOC lt SOChigh thevehicle is in hybrid traction mode B Considering the effi-ciency of battery and the fuel economy of engine the powerdistribution between the battery and IGPU is carried out bya Mamdani fuzzy logic controller (FLC) with double inputsand single outputThe IGPU load coefficient 120582 (120582 = 119875119903119890119902260)and battery SOC are selected as input variables powerdistribution coefficient 119896119892 is selected as output variable Thetarget power of the battery and IGPU can be expressedas

1198751015840119887 = (1 minus 119896119892) sdot 1198751199031198901199021198751015840119892 = 119896119892 sdot 119875119903119890119902 (5)

The relative membership function diagram is shown inFigure 5 120582 SOC and 119896119892 all have five fuzzy subsets VSMS M MB VB and the basic domain of 120582 SOC and 119896119892respectively are [01 12] [05 08] and [07 15] In the designprocess of FLC the following principles should be notedWhen 120582 is small 119896119892 should be raised so as to make the engineoperate in its optimal operation region When 120582 is large kgshould be reduced so that the battery can supplement theadditional power When SOC is large or small 119896119892 should bereduced or raised so that the battery can be discharged orcharged in time to keep SOC in the area with high efficiencyBy using the rule formof ldquoif120582 isA and SOC isB then 119896119892 isCrdquo25 fuzzy rules are established as shown in Table 2

34 Braking Mode Under driving mode 119875119903119890119902 can be calcu-lated according to the angular displacement of acceleratorpedal However for braking mode 119875119903119890119902 cannot be calculatedby using the angular displacement of brake pedal becausethe electrical regenerative braking coexists with mechanicalbraking In this paper the braking force of the motors is


VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

1203 04 05 06 07 08 09 1 110201

0

05

1

Deg

ree o

f mem

bers

hip

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

08 09 1 11 12 13 14 1507Kg

Figure 5 Membership function of the FLC

Table 2 Fuzzy control rules

120582 SOCVS MS M MB VB

VS VB VB MS MS MSMS VB MB MS MS VSM MB MB M MS VSMB MB M M MS VSVB MS VS VS VS VS

calculated by using of I curve and then 119875119903119890119902 is obtainedcombining with the vehicle speed I curve is an ideal brakingforce distribution curve as shown in Figure 6 where j isthe deceleration of the vehicle during braking Under thiscondition the front axle braking force 1198651205831 and rear axlebraking force 1198651205832 are satisfied with the following formula

1198651205832 = 12 [[

119866ℎ119892radic1198872 + 4ℎ119892119871119866 1198651205831 minus (119866119887ℎ119892 + 21198651205831)]]

(6)

where G is the vehicle gravity ℎ119892 is the height of vehiclecentroid b is the distance between vehicle centroid and rearaxle and L is the wheelbase

Besides the sum of 1198651205831 and 1198651205832 is equal to the overallbraking force that is

120573 sdot 119865119887119903119886 max = 1198651205831 + 1198651205832 (7)

where 120573 is conversion coefficient of brake pedal angledisplacement (120573 isin [minus1 0]) and 119865119887119903119886 max is the maximum

01 02 03 04 05 06 07 08 100

01

02

03

04

05

06

09

a

b

I curve

j=10gj=09gj=08gj=07gj=06gj=05gj=04g

j=03gj=02g

j=01g

Rear

bra

king

forc

e rat

io(F

2G)

Ffr2

FmＧＲ

FmＧＲ

Freg1

Freg2

Ffr1 Front braking force ratio (F1G)

Figure 6 Ideal braking force distribution curve

braking force of the vehicle Combining formula (6) andformula (7)1198651205831 and 1198651205832 can be obtained 1198651205831 and 1198651205832 are usu-ally composed of mechanical braking force and regenerativebraking force For more energy recapture the regenerativebraking should be given priority If 1198651205831 (1198651205832) is smaller thanthe maximum regenerative braking force of motor the totalbraking force is provided by the motor and the mechanicalbrake is not used as shown by point a in Figure 6 If 1198651205831(1198651205832) is greater than the maximum regenerative braking forceof motor the motor is controlled to produce the maximumregenerative braking force and the remaining force1198651198911199031 (1198651198911199032)is provided by mechanical brake as shown by point b inFigure 6 that is

1198651199031198901198921 = minusmin (1198651205831 119865119898max)1198651199031198901198922 = minusmin (1198651205832 119865119898max) (8)

where1198651199031198901198921 1198651199031198901198922 are the target regenerative braking forceon the front axle and rear axle respectively 119865119898max is themaximum regenerative force of motor and its value is limitedby the motor speedN In addition it should be noted that theregenerative braking is affected by braking strength Z vehiclespeed V and battery SOC

When 0ltZlt07 the motor can be controlled for regen-erative braking When Zge07 the vehicle is in an emergencybraking state Considering braking safety the motor shouldnot work and the mechanical brake has to produce the totalbraking force as required The braking strength correctionfactor 1198961 is introduced as

1198961 = 1 0 le 119885 lt 070 119885 ge 07 (9)

When Vle12kmh it is hard for the motors to producebrake torque due to the low speedTherefore with the vehiclespeed drops to 12 kmh the regenerative braking force should


be gently reduced to 0 The speed correction factor 1198962 isintroduced as

1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)

When the value of SOC is very high (SOCgt099) thebattery will be damaged if it is still charged Therefore withthe battery SOC up to 099 the regenerative braking forceshould also be gently reduced to 0The SOC correction factor1198963 is introduced as

1198963 =

1 SOC lt 08minus6667SOC + 6333 08 le SOC lt 0990 SOC ge 099

(11)

The actual regenerative braking force on the front axle andrear axle 11986510158401199031198901198921 and 11986510158401199031198901198922 can be expressed as

11986510158401199031198901198921 = 1198961 sdot 1198962 sdot 1198963 sdot 119865119903119890119892111986510158401199031198901198922 = 1198961 sdot 1198962 sdot 1198963 sdot 1198651199031198901198922 (12)

Therefore the 119875119903119890119902 under braking state can be expressedas

119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)

where r is wheel radius and 1198940 is final drive ratio Thebraking state can be divided into two cases according to SOCvalue (module 7 and module 8 in Figure 4) When SOCltSOClow the vehicle is in braking mode AThe battery shouldbe charged with maximum current and the brake resistancedoes not work that is

1198751015840119887 = minusmin (1003816100381610038161003816119875119888max1003816100381610038161003816 119875119892max minus 119870119887119903119886119875119903119890119902)

1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)

where 119870119887119903119886 is the maximum regenerative braking coeffi-cient When SOCge SOClow the vehicle is in braking mode

BThe battery is charged normally and IPGU does not workIf the recovery energy is too much the excess energy can beconsumed in the form of thermal energy by using the brakeresistance that is

1198751015840119887 = minusmin (1003816100381610038161003816119875119888max1003816100381610038161003816 minus119870119887119903119886119875119903119890119902)

1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)

4 Implementation of the IQGA for FLC

Because the FLC design is mainly based on the engineeringexperiences of experts there is room for improvement In thissection in order to improve the fuel economyof hybrid powersystem the IQGA is proposed to optimize the fuzzy controlstrategy

41 Improved Quantum Genetic Algorithm The quantumgenetic algorithm (QGA) is a probability search algorithmbased on the theory of quantum computing which isproposed by Han [34] In QGA quantum state vector isused for genetic encoding and qubit rotation gate is usedto realize chromosome evolution Therefore the algorithmhas strong parallelism which can achieve better calculationresults than conventional genetic algorithm (GA) Howeverits convergence performance leaves room to improve TheIQGA is proposed in this paper based on the improvementof QGA The dynamic adjusting rotation gate is adopted toimprove the convergence speed and quantum catastrophicoperation is introduced to overcome premature convergenceand improve global optimization ability

411 Qubit Encode A qubit can be interpreted not only ldquo0rdquostate or ldquo1rdquo state but also any superposition of them It can beexpressed as

1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)

where 120572 and 120573 are complex numbers satisfying |120572|2 +|120573|2 = 1 |120572|2 and |120573|2 represent the probability of qubitin 0 state and 1 state respectively The multiqubits can beused to represent the chromosome with multigenes as fol-lows

119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512

1003816100381610038161003816100381610038161003816100381610038161003816sdot sdot sdotsdot sdot sdot

100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)

where 119902119905119895 is the j-th individual chromosome in t-thgeneration k is the number of qubits used for each geneencoding and m is the number of genes in the chromo-some

412 Measure The measuring method for the t-th genera-tion population 119876(119905) = (1199021199051 1199021199052 119902119905119899) is as follows For eachqubit generate a random number Nrand between 0 and 1 ifNrand le |120572|2 the qubit is 0 otherwise 1 Then the binary


Table 3 Look-up table of function119891(120572119894 120573119894)119887119894 119861119890119904119905119894 119891(119909119894) gt 119891(119861119890119904119905) Δ120579119894 119891(120572119894 120573119894)120572119894 sdot 120573119894 gt 0 120572119894 sdot 120573119894 lt 0 120572119894 = 0 120573119894 = 00 0 False 0 0 0 0 00 0 True 0 0 0 0 00 1 False 120575 1 -1 0 plusmn10 1 True 120575 -1 1 plusmn1 01 0 False 120575 -1 1 plusmn1 01 0 True 120575 1 -1 0 plusmn11 1 False 0 0 0 0 01 1 True 0 0 0 0 0

coding population 119875(119905) = (1199091199051 1199091199052 119909119905119899) can be obtainedSuppose 119883119898 represents the decimal value corresponding tothe m-th gene 119909119898 = (119887119896119887119896minus1 sdot sdot sdot 11988721198871) which is in the rangebetween 119880max

119898 and 119880min119898 119909119898 can be converted to 119883119898 as

follows

119883119898 = 119880min119898 + ( 119896sum

119894=1

119887119894 sdot 2119894minus1) sdot 119880max119898 minus 119880min

1198982119896 minus 1 (18)

Then the decimal measurement value of the 119899th chromo-some 119883119905119899 = (1198831 1198832 sdot sdot sdot 119883119898) can be calculated By the samemethod the decimal value of the t-th generation population119878(119905) = (1198831199051 1198831199052 sdot sdot sdot 119883119905119899) is obtained413 Qubit Rotation Gate Strategy The quantum rotationgate 119880(120579) is the executing mechanism of updating operationwhich directly affects the algorithm performance It can beexpressed as

119880 (120579) = [cos (120579) minussin (120579)sin (120579) cos (120579) ] (19)

The updating operation of qubit is as follows

[12057210158401198941205731015840119894] = 119880 (120579) [120572119894120573119894] = [120572119894 cos (120579) minus 120573119894 sin (120579)120572119894 sin (120579) + 120573119894 cos (120579)] (20)

where (120572119894 120573119894)119879 and (1205721015840119894 1205731015840119894 )119879 are the probability amplitudeof 119894th qubit before and after updating respectively 120579 is therotation angle in general 120579 isin [0001120587 005120587] and it can beexpressed as

120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)

where Δ120579119894 and 119891(120572119894 120573119894) are the value and direction ofthe rotation angle and they can be obtained by Table 3In Table 3 119887119894 and 119861119890119904119905119894 respectively are the 119894th bit of thecurrent measuring individual and optimal individual 119891(119909119894)and 119891(119861119890119904119905) respectively are the fitness values of the currentmeasuring individual and optimal individual 120575 is a dynamicvalue which is used to control the convergence speed of thealgorithm it can be expressed as

120575 = 120579min

+ (119890(119872119860119883119866119864119873minus119905)(119872119860119883119866119864119873minus1) minus 1119890 minus 1 ) (120579max minus 120579min) (22)

where t is the evolution generation 119872119860119883119866119864119873 is thelargest evolution generation and 120579max and 120579min are themaximum rotation angle and minimum rotation anglerespectively 120575 is decreased with the increase of evolutiongeneration so as to realize the dynamic adaptive adjustmentof the qubit rotation gate

414 Population Catastrophe After several generations ofevolution the algorithm may fall into the local optimalsolution The population catastrophe can be used to solvethis problem For each evolution the average fitness of thepopulation 119891119886V119892 can be calculated as follows

119891119886V119892 = sum119899119894=1 119891119894119899 (23)

where 119891119894 represents the fitness value of the 119894th individualand n is the number of individuals in the population Thefitness values which is greater than 119891119886V119892 are averaged to get1198911015840119886V119892 and define the catastrophe factor Δ119891 as

Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)

If Δ119891 is very small it indicates that the diversity of thepopulation is poor and the catastrophe should be carried outat this time that is only the best individual is retained andthe others should be regenerated

42 FLC Optimization The membership function is the keyfactor affecting the performance of FLC so the decisiveparameters of the membership function are optimized inthis paper Take 120582 as an example as shown in Figure 7 thelocation of the vertexes of the red triangles A B and Cdetermines the shape of the membership functionThereforethe distance between the line segments AD AE and EC (iex1 x2 and x3) can be used as the variables to be optimizedThe number of variables can be reduced effectively by thismethodwhich greatly reduces the computational complexityThe same coding method is used to optimize the membershipfunction of SOC and 119896119892 Thus there are 9 variables to beoptimized that is there are 9 genes in the chromosome

During the driving cycle both the battery and the engine-generator set need to provide energy for the vehicle Toeliminate the effect of battery energy on fuel consumption


VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip

Figure 7 Optimized variables of 120582

the battery SOC of final state before and after optimizationshould be controlled at the same level The ideal situation isthat the final state SOC difference ΔSOC is controlled to 0but in this case the amount of calculation will become verylarge In order to improve the speed of simulation a certaintolerance of ΔSOC can be set In this paper the tolerance is05 [35 36] that is the constraint on SOC is defined as

minus05 le ΔSOC le 05 (25)

Under the condition of satisfying vehicle power per-formance aiming at the goal of low fuel consumption inurban and high speed driving cycle the objective function isobtained as follows [37]

119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)

where 119862119894119905119910 119881119887 and 119867119908119910 119881119887 are the engine fuel con-sumption (L) under the urban and high speed driving cyclerespectively

The optimization process of the fuzzy controller basedon IQGA is shown in Figure 8 and the specific steps are asfollows(1) Initialize the population 119876(1199050) randomly generate nchromosomes encoded with qubits(2) Get the confirmed solution 119878(1199050) by measuring 119876(1199050)(3) Assign 119878(1199050) to FLC evaluate 119876(1199050) and record theoptimal individual 1199021199050

119887119890119904119905and its fitness 1198741198871198951198651199061198991199050

119887119890119886119905(4) Decide whether the calculation process can be fin-

ished if it meets the termination condition end or go on(5) Get the confirmed solution 119878(119905) by measuring 119876(119905)assign 119878(119905) to FLC evaluate 119876(119905) and record the individualfitness 119874119887119895119865119906119899119905119894 (119894 isin 1 119899) of 119876(119905)(6) Update119876(119905) by using quantum rotation gate 119880(120579) andpopulation catastrophe get new population 119876(119905 + 1)(7) Evaluate 119876(119905 + 1) get the optimal individual 119902119905119887119890119904119905 andits fitness 119874119887119895119865119906119899119905119887119890119886119905(8) t + 1 return to step (4)5 Simulation and HILS Experiment

In this section the effectiveness of the proposed EMS forseries hybrid electric rescue vehicle is validated Firstly thesimulation model of the vehicle is built with Simulink Thenthe offline optimization simulation is carried out Finally

Start

Meet the terminatingconditions

EndObtain S(t) by measuring Q(t)

Assign S(t) to FLC evaluate Q(t)

t=t+1

YES

NO

Initialize population Q(t0)

Obtain S(t0) by measuring Q(t0)

Assign S(t0) to FLC evaluate Q(t0)

Obtain qt0best

ObjFunt0beat

Obtain ObjFunti

Get Q(t+1) by using U() and population catastrophe

Evaluate Q(t+1) Obtain qtbest ObjFuntbeat

Figure 8 FLC optimization process based on IQGA

using the DISPACE platform the driver-controller basedHILS platform is established with which to verify the real-time performance of the control strategy

51 System Mathematical Models The simulation model ofseries hybrid electric rescue vehicle as shown in Figure 9 ismainly made up of driver model drive torque distributionmodel braking torque distribution model vehicle dynam-ics model control strategy model driving mode judgmentmodel motor model engine-generator model DCDC con-verter model and battery model The driver model is builtbased on PID control and the error of target speed andactual speed is taken as the input [38] The driving torquedistributionmodel braking torque distributionmodelmotormodel and control strategy model are built on the basis of the


Figure 9 Simulation model of series hybrid electric rescue vehicle

EMS proposed in Section 3The following mainly introducesthe vehicle dynamics model engine-generator model andbattery model

(1) Vehicle Dynamics Model The vehicle dynamics model isbuilt based on longitudinal dynamics characteristics of thevehicle the torque of wheel 119879119908 can be expressed as

119879119908 = 1198940 sdot (1205781198791 sdot 1198791 + 1205781198792 sdot 1198792) minus (119879119898119890119888ℎ1 + 119879119898119890119888ℎ2) (27)

where 1205781198791 and 1205781198792 are the transmission efficiency of thefront and rear motor respectively 1198791 and 1198792 are the outputtorque of the front and rear motor respectively 119879119898119890119888ℎ1 and119879119898119890119888ℎ2 are the mechanical braking torque of the front and rearaxle respectively 119879119908 can also be expressed as

119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)

where m is the vehicle mass (kg) g is the gravityacceleration Cr is the rolling resistance coefficient 120572 is theroad-grade angle 119862119863 is the air drag coefficient 120588 is the airdensity A is the frontal area of the vehicle (m2) 1205751015840 is thecorrection coefficient of rotating mass and u is the vehiclespeed (ms) The front and rear motor speed N1 and N2(rmin) have the following relationship with u

1198731 = 1198732 = 301198940119906120587119903 (29)

(2) Engine-Generator Model Because there is velocity cou-pling between the engine and generator they are modeledas a whole The equivalent circuit of engine-generator setis shown in Figure 10 [39] Ignoring the internal resistanceand torque loss the DC bus voltage Udc and the generatorelectromagnetic torque Tg can be expressed as

+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc

Figure 10 The equivalent circuit of engine-generator set

119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)

where Ke and Kx are the electromotive force coeffi-cient and equivalent impedance coefficient of the generatorrespectively and they can be obtained from the bench testdata 120596g is the rotation angular velocity of generator Idc is theDC bus current The engine and the generator are connectedby a speed increase box and the speed increasing ratio is iegAccording to the torque balance the following equations canbe obtained

119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)

where Teng is the engine torque and Je and Jg are themoment of inertia of engine and generator respectively nengand ng are the rotational speed of engine and generator


+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib

Figure 11 The equivalent circuit of battery

respectively In addition the engine fuel consumption Ve canbe calculated from formula (32)

119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)

where be is the fuel consumption rate (be=f (Teng neng))120588119890 is the fuel density and 120578119890 is the engine output efficiency

(3) Battery Model The battery model is built by using thetypical Rint model [40] and the equivalent circuit is shownin Figure 11 Ub V0 Ib Rch and Rdch are the terminalvoltage open-circuit voltage internal current (positive whendischarge) charge resistance and discharge resistance of thebattery respectively Among them the values of V0 Rchand Rdch are all related to the battery SOC so the followingequations can be obtained

119880119887 = 1198810 minus 119868119887119877119889119888ℎ 119868119887 gt 0119880119887 = 1198810 minus 119868119887119877119888ℎ 119868119887 lt 01198810 = 1198911 (SOC)

119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)

The ampere hour method [41] is applied to estimate thebattery SOC Suppose the maximum capacity of the batteryis 119876max and the initial capacity of the battery is Q0 then thebattery SOC at time t can be expressed as

119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)

52 Simulation The driving cycle used in the simulation asshown in Figure 12 ismainly composed of UDDSurban driv-ing cycle and HWFET high speed driving cycle In order tocompare the optimization effect of different algorithms GAQGA and IQGA are used to optimize the FLC respectivelySupposing that the genetic algebra is 200 that the populationsize is 50 and that the binary length of each variable is 10 thesimulation is carried out by using MATLABSimulink The

500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)

Figure 12 UDDS and HWFET combined driving cycle

20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA

Figure 13 Optimization process of different algorithms

changing curves of global fitness with different strategies areshown in Figure 13 It can be seen that under the conditionof same population size and generation the fitness of IQGAreaches the minimum value 38 after 80 generations whilethe fitness of QGA reaches the minimum value 385 after 114generations and the fitness of GA reaches the minimum value402 after the 104 generations It shows that IQGA has fasterconvergence speed and better optimization result than QGAand GA which is more suitable for solving the optimizationproblem of EMS The fuzzy control membership functionoptimized by IQGA is shown in Figure 14

Figures 15 and 16 respectively reveal the distribution ofIGPU operating points before and after IQGA optimizationIt can be seen that compared with the original fuzzy controlstrategy the optimized IGPU operating points are moredistributed in the high efficiency regionThe changing curvesof battery SOC before and after optimization are shown inFigure 17 It shows that the initial SOC value is 07 afterIQGA optimization the final SOC value changed from 0677to 0674 It is observed that Δ119878119874119862= -044 which satisfiesthe SOC constraints set in (25) Figure 18 shows the fuelconsumption before and after optimization It can be seenthat the fuel consumption is 871L before optimization whileit drops by 517 to 826L after optimization Therefore theresults indicate that with IQGA optimized control strategy


VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip

Figure 14 Membership function optimized by IQGA

Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)

Fuel consumption rate (gkWh)Maximum output power curveOptimum operating lineOperating points

Figure 15 IGPU operating points before IQGA optimization

the fuel consumption can be effectively reduced and the IGPUefficiency can be significantly improved

53 HILS Experiment In offline simulation the driverrsquosoperation habits and the real-time operation effect of controlstrategy are not considered In order to validate the proposedEMS in real-world implementation the HILS experimentbased on driver and controller is carried out with dSPACE

Using dSPACE the Simulink models could be convertedinto C codes and run reliably [42 43]The schematic diagram

Engine speed (rmin)800 1000 1200 1400 1600 1800 2000

50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308

Figure 16 IGPU operating points after IQGA optimization

500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC

Original FLCIQGA optimized FLC

Figure 17 Changing curve of battery SOC

0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)


Figure 18 Fuel consumption


Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver

Engine-generator Motor1

CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer

Vehicle dynamics model

Operating signal

General controller

BatteryMotor2

Figure 19 Schematic diagram of HILS platform

dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer

Figure 20 Hardware of HILS platform

of HILS platform is shown in Figure 19 The driver controlsthe acceleration or brake pedal to produce analog signalswhich are transmitted to the general controller through theAD interface After calculation and analysis the controlsignal produced by general controller is input to dSPACEthrough the CAN bus According to the control instructionthe engine-generator model battery model motor modeland vehicle dynamics model are run in dSPACE Meanwhilethe operating parameters are displayed on the monitor Thehost computer is used to modify compile and download theinternal models of dSPACE The hardware of HILS platformis shown in Figure 20

According to the driverrsquos driving habits the HILS exper-iment is carried out under a typical 70 seconds drivingcondition and the results are shown in Figure 21 Figure 21(a)shows the driverrsquos operation signal and Figure 21(b) shows

the vehicle speed As we can see in 1-11s the vehicleexperienced a slight acceleration and the speed increasedfrom 0 to 11 kmh In 12-17s the vehicle experienced a slightbraking and the speed fell to 2 kmh In 18-33s the vehicleexperienced a medium strength acceleration and the speedrose to 47 kmh In 34-41s the vehicle experienced a mediumstrength braking and the speed fell to 14 kmh In 42-62sthe vehicle experienced two types of quick acceleration andthe speed rose to 88 kmh In 63-65s and 69-70s the vehiclewas in the state of gliding and the speed decreased forthe driving resistance In 66-68s the vehicle experienced anemergency braking and the speed fell to 3 kmh It can beseen that the vehicle speed curve is smooth and stable whichcan follow the driverrsquos acceleration or braking intention verywell

Figure 21(c) shows the motor demand power 119875119903119890119902 IGPUoutput power 119875119892 and battery output power 119875119887 It can be seenthat the battery plays a good role in reducing peak and fillingvalley when the demand power is negative the energy canbe recovered effectively when the demand power is high thecompensation power can be output in time Meanwhile theengine works in the efficient area above 50KW and when thedemand power is small it can be properly adjusted to outputlarger power to charge the battery

In Figure 21(d) the torque distribution of the front andrear axle is shown When the vehicle is driven the drivingtorque on the front axle is the same with that of rear axleWhen the vehicle is braking the braking torque on thefront and rear axle can be distributed according to the Icurve

Figures 21(e) and 21(f) show the distribution of mechan-ical torque and motor torque on front and rear axle respec-tively It can be seen that when the vehicle is driven the motortorque can basically meet the demand torque but whenthe vehicle speed is high the motor torque is not sufficientenough for accelerating When the vehicle is braking it needs


10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)

Target torqueFront axle torqueRear axle torque

minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)

Front axle torqueMotor torqueMechanical torque

minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)

Rear axle torqueMotor torqueMechanical torque

minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)

Figure 21 The results of HILS experiment

to be analyzed in different situations In 12-17s because thespeed is lower than 12 kmh only mechanical braking iscarried out In 34-38s the braking strength is less than 07and the require torque is smaller than the maximum brakingtorque ofmotor so only regenerative braking is carried out In39-41s the speed is reduced from 24 kmh to 14 kmh so theregenerative braking torque gradually becomes smaller andthe mechanical braking torque gradually becomes larger In66-68s the vehicle is in an emergency braking state and thebraking strength is more than 07 so the braking torque istotally provided by the mechanical brake

6 Conclusions

In this paper an energy management strategy combiningfuzzy control and threshold control was proposed for aseries hybrid electric rescue vehicle and a FLC optimizationmethod based on IQGA was introduced to further improvethe fuel economy The effectiveness and real-time perfor-mance of the proposed EMS were validated by simulationand HILS experiment The simulation results show thatcompared with QGA and GA the IQGA optimization algo-rithm has faster convergence speed and better optimization


result With IQGA optimization EMS the engine efficiencyis obviously improved and the fuel consumption can bereduced by 517 compared with that before optimizationTheHILS experiment shows that with the proposed EMS thereal-time requirements of the vehicle can be well satisfied theworking characteristics of IGPUand battery can be effectivelyexploited and the braking energy can be fully absorbed by thefront and rearmotor In conclusion the EMS proposed in thispaper has achieved good control effect which can be used inreal-time control of the series hybrid electric rescue vehicle

Data Availability

The data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

This work was financially supported by National Key RampDProgram of China (2016YFC0802900)

References

[1] K C Bayindir M A Gozukucuk and A Teke ldquoA com-prehensive overview of hybrid electric vehicle powertrainconfigurations powertrain control techniques and electroniccontrol unitsrdquo Energy Conversion and Management vol 52 no2 pp 1305ndash1313 2011

[2] Ehsani Mehrdad Gao Yimin and Ali EmadiModern ElectricHybrid Electric And Fuel Cell Vehicles Fundamentals TheoryAnd Design CRC Press Boca Raton Fl USA 2009

[3] C Dextreit and I V Kolmanovsky ldquoGame theory controller forhybrid electric vehiclesrdquo IEEE Transactions on Control SystemsTechnology vol 22 no 2 pp 653ndash663 2014

[4] H Banvait S Anwar and Y Chen ldquoA rule-based energy man-agement strategy for Plugin Hybrid Electric Vehicle (PHEV)rdquoin Proceedings of the American Control Conference (ACC rsquo09)pp 3938ndash3943 St Louis Mo USA June 2009

[5] X Wang H He F Sun and J Zhang ldquoApplication study onthe dynamic programming algorithm for energy managementof plug-in hybrid electric vehiclesrdquo Energies vol 8 no 4 pp3225ndash3244 2015

[6] B C Chen Y Y Wu and H C Tsai ldquoDesign and analysis ofpower management strategy for range extended electric vehicleusing dynamic programmingrdquo Applied Energy vol 113 pp1764ndash1774 2014

[7] C Hou M Ouyang L Xu and H Wang ldquoApproximate Pon-tryaginrsquosminimumprinciple applied to the energymanagementof plug-in hybrid electric vehiclesrdquo Applied Energy vol 115 pp174ndash189 2014

[8] S Zhang R Xiong and C Zhang ldquoPontryaginrsquos MinimumPrinciple-based power management of a dual-motor-drivenelectric busrdquo Applied Energy vol 159 pp 370ndash380 2015

[9] Z Zhou and C Mi ldquoPower management of PHEV usingquadratic programmingrdquo International Journal of Electric andHybrid Vehicles vol 3 no 3 pp 246ndash258 2011

[10] Z Chen C C Mi R Xiong J Xu and C You ldquoEnergy man-agement of a power-split plug-in hybrid electric vehicle basedon genetic algorithm and quadratic programmingrdquo Journal ofPower Sources vol 248 pp 416ndash426 2014

[11] Z Yarning S Fengchun and H Hongwen ldquoControl strategyoptimization for hybrid electric vehicle based on DIRECTalgorithmrdquo in Proceedings of the 2008 IEEE Vehicle Power andPropulsion Conference VPPC 2008 China September 2008

[12] A Panday and H O Bansal ldquoFuel efficiency optimization ofinput-split hybrid electric vehicle using DIRECT algorithmrdquoin Proceedings of the 9th IEEE International Conference onIndustrial and Information Systems ICIIS 2014 IndiaDecember2014

[13] J Hao Z Yu Z Zhao P Shen and X Zhan ldquoOptimizationof key parameters of energy management strategy for hybridelectric vehicle using DIRECT algorithmrdquo Energies vol 9 no12 2016

[14] P Elbert T Nuesch A Ritter N Murgovski and L GuzzellaldquoEngine onoff control for the energy management of a serialhybrid electric bus via convex optimizationrdquo IEEE Transactionson Vehicular Technology vol 63 no 8 pp 3549ndash3559 2014

[15] J Park and J-H Park ldquoDevelopment of equivalent fuel con-sumption minimization strategy for hybrid electric vehiclesrdquoInternational Journal of Automotive Technology vol 13 no 5pp 835ndash843 2012

[16] T Nuesch A Cerofolini and G Mancini ldquoEquivalent con-sumption minimization strategy for the control of real drivingNOx emissions of a diesel hybrid electric vehiclerdquo Energies vol7 no 5 pp 3148ndash3178 2014

[17] G Paganelli T M Guerra S Delprat J J Santin M Delhomand E Combes ldquoSimulation and assessment of power controlstrategies for a parallel hybrid carrdquo Proceedings of the InstitutionofMechanical Engineers Part D Journal of Automobile Engineer-ing vol 214 no 7 pp 705ndash718 2000

[18] K Yu H Yang X Tan et al ldquoModel Predictive Control forHybrid Electric Vehicle Platooning Using Slope InformationrdquoIEEE Transactions on Intelligent Transportation Systems vol 17no 7 pp 1894ndash1909 2016

[19] X Zeng and JWang ldquoA Parallel Hybrid Electric Vehicle EnergyManagement Strategy Using Stochastic Model Predictive Con-trol With Road Grade Previewrdquo IEEE Transactions on ControlSystems Technology vol 23 no 6 pp 2416ndash2423 2015

[20] V T Minh and A A Rashid ldquoModeling and model predictivecontrol for hybrid electric vehiclesrdquo International Journal ofAutomotive Technology vol 13 no 3 pp 477ndash485 2012

[21] KM Passino and S Yurkovich ldquoFuzzy Controlrdquo inTheControlSystems Handbook CRC Press 2001

[22] L Wang ldquoStable adaptive fuzzy control of nonlinear systemsrdquoIEEE Transactions on Fuzzy Systems vol 1 no 2 pp 146ndash1551993

[23] S G Li SM Sharkh F CWalsh andCN Zhang ldquoEnergy andbattery management of a plug-in series hybrid electric vehicleusing fuzzy logicrdquo IEEE Transactions on Vehicular Technologyvol 60 no 8 pp 3571ndash3585 2011

[24] H Hannoun D Diallo and C Marchand ldquoEnergy manage-ment strategy for a parallel hybrid electric vehicle using fuzzylogicrdquo in Proceedings of the International Symposium on PowerElectronics Electrical Drives Automation and Motion 2006SPEEDAM 2006 pp 230ndash234 Italy May 2006


[25] A S Sarvestani and A A Safavi ldquoA novel optimal energymanagement strategy based on fuzzy logic for a hybrid electricvehiclerdquo in Proceedings of the 2009 IEEE International Confer-ence on Vehicular Electronics and Safety ICVES 2009 pp 141ndash145 India November 2009

[26] D Zhao R Stobart G Dong and E Winward ldquoReal-TimeEnergy Management for Diesel Heavy Duty Hybrid ElectricVehiclesrdquo IEEETransactions on Control Systems Technology vol23 no 3 pp 829ndash841 2015

[27] Y Yang P Ye X Hu B Pu L Hong and K Zhang ldquoAresearch on the fuzzy control strategy for a speed-couplingISGHEV based on dynamic programming optimizationrdquoQicheGongchengAutomotive Engineering vol 38 no 6 pp 674ndash6792016

[28] A Poursamad and M Montazeri ldquoDesign of genetic-fuzzycontrol strategy for parallel hybrid electric vehiclesrdquo ControlEngineering Practice vol 16 no 7 pp 861ndash873 2008

[29] J Liang J Zhang H Zhang and C Yin ldquoFuzzy energy man-agement optimization for a parallel hybrid electric vehicle usingchaotic non-dominated sorting genetic algorithmrdquo Automatikandash Journal for Control Measurement Electronics Computing andCommunications vol 56 no 2 pp 149ndash163 2015

[30] J Wu C-H Zhang and N-X Cui ldquoFuzzy energymanagementstrategy for a hybrid electric vehicle based on driving cyclerecognitionrdquo International Journal of Automotive Technologyvol 13 no 7 pp 1159ndash1167 2012

[31] N Limin H Yang and Y Zhang ldquoIntelligent HEV FuzzyLogic Control Strategy Based on Identification and Predictionof Drive Cycle andDriving TrendrdquoWorld Journal of Engineeringamp Technology vol 3 no 3 pp 215ndash226 2015

[32] Y L Murphey Z Chen L Kiliaris and M A Masrur ldquoIntel-ligent power management in a vehicular system with multiplepower sourcesrdquo Journal of Power Sources vol 196 no 2 pp 835ndash846 2011

[33] R Langari and J-SWon ldquoIntelligent energymanagement agentfor a parallel hybrid vehiclemdashpart I system architecture anddesign of the driving situation identification processrdquo IEEETransactions on Vehicular Technology vol 54 no 3 pp 925ndash934 2005

[34] K H Han and J H Kim ldquoGenetic quantum algorithm andits application to combinatorial optimization problemrdquo in Pro-ceedings of the IEEE International Conference on EvolutionaryComputation pp 1354ndash1360 San Diego Calif USA July 2000

[35] S Borthakur and S C Subramanian ldquoParameter matchingand optimization of a series hybrid electric vehicle powertrainsystemrdquo in Proceedings of the ASME 2016 International Mechan-ical Engineering Congress and Exposition IMECE 2016 USANovember 2016

[36] Q Tu X Zhang M Pan C Jiang and J Xue ldquoDesign andOptimization of the Power Management Strategy of an ElectricDrive Tracked VehiclerdquoMathematical Problems in Engineeringvol 2016 Article ID 6185743 13 pages 2016

[37] K Wipke T Keith and D Nelson ldquoOptimizing EnergyManagement Strategy and Degree of Hybridization for aHydrogen Fuel Cell SUVrdquo httpswwwresearchgatenetpubli-cation2868610 Optimizing Energy Management Strategy andDegree of Hybridization for a Hydrogen Fuel Cell SUV

[38] T A T Mohd M K Hassan I Aris C S Azura and B S KK Ibrahim ldquoApplication of fuzzy logic in multi-mode drivingfor a battery electric vehicle energymanagementrdquo InternationalJournal on Advanced Science Engineering and InformationTechnology vol 7 no 1 pp 284ndash290 2017

[39] F Sun and C Zhang Technoloyies for the hybrid electric drivesystem of armored vehicle NationalDefense Industry Press 2017

[40] HW He R Xiong H Q Guo and S C Li ldquoComparison studyon the battery models used for the energy management of bat-teries in electric vehiclesrdquo Energy Conversion and Managementvol 64 pp 113ndash121 2012

[41] W Waag C Fleischer and D U Sauer ldquoOn-line estimation oflithium-ion battery impedance parameters using a novel varied-parameters approachrdquo Journal of Power Sources vol 237 pp260ndash269 2013

[42] K Meah S Hietpas and S Ula ldquoRapid control prototyping ofa permanent magnet DC motor drive system using dSPACEand mathworks simulinkrdquo in Proceedings of the APEC 2007 -22nd Annual IEEE Applied Power Electronics Conference andExposition pp 856ndash861 USA March 2007

[43] A Rubaai and P Young ldquoHardwaresoftware implementationof fuzzy-neural-network self-learning control methods forbrushless DC motor drivesrdquo IEEE Transactions on IndustryApplications vol 52 no 1 pp 414ndash424 2016

Hindawiwwwhindawicom Volume 2018

MathematicsJournal of


Mathematical Problems in Engineering

Applied MathematicsJournal of


Probability and StatisticsHindawiwwwhindawicom Volume 2018

Journal of


Mathematical PhysicsAdvances in

Complex AnalysisJournal of


OptimizationJournal of



Engineering Mathematics

International Journal of


Operations ResearchAdvances in

Journal of


Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018

International Journal of Mathematics and Mathematical Sciences


Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018Volume 2018

Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in

Nature and SocietyHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Dierential EquationsInternational Journal of

Volume 2018


Decision SciencesAdvances in


AnalysisInternational Journal of


Stochastic AnalysisInternational Journal of

Submit your manuscripts atwwwhindawicom


Motorcontroller

General controller

DC bus

CAN busElectric linkMechanical link

Brakingcontroller

Motor

Brake

Brake

Brake

Brake

Finaldrive MotorIGPU

controller

ACDC

IGPU

BMS

Battery pack

DCDC

Motorcontroller

Finaldrive

Figure 1 Structure diagram of the hybrid rescue vehicle

such as fuel consumptionminimization strategy (ECMS) [15ndash17] and model predictive control strategy (MPC) [18ndash20]However the instantaneous optimization belongs to localoptimization and the optimization effect depends on theprecision of the mathematical model

(3) Fuzzy Logic Control Strategy With good real-time per-formance and robustness [21 22] the fuzzy logic controlstrategy does not rely on precisemathematical models whichis suitable for dealing with the complex nonlinear andtime-varying problems In practical applications the speedSOC required torque and power are often used as inputvariables [23ndash25] and the power ratio of the engine andthe battery is used as the output variable However theformulation of control logic is mainly based on engineeringexperience and it is difficult to ensure the optimal powerdistribution To achieve better performance many strategiescombining fuzzy control method with other control methodshave been proposed by some scholars In [26] a fuzzycontroller was used to adjust the equivalent factors of ECMSand the fuel economy of the hybrid electric heavy-dutyvehicle was improved In [27] the DP algorithm was usedto optimize the performance of ISG hybrid electric vehicleand combined with the optimization results a more practicalfuzzy controller was proposed In [28ndash30] some intelligentalgorithms such as genetic algorithm and particle swarmoptimization were used to optimize the membership func-tion or control rules of the fuzzy controller In [31ndash33] themixed-fuzzy control strategy was designed by using neuralnetwork working condition identificationmachine learningand other intelligent techniques However the control effectsof the above strategies still exist a certain improvement space

Considering the advantages and disadvantages of theabove three categories of methods a novel EMS is proposed

to improve the fuel economy of a series hybrid electricrescue vehicle in this paper First combined with thresholdcontrol and fuzzy control the EMS is initially designedaccording the working characteristics of engine and batteryThen an improved quantum genetic algorithm with betterconvergence performance is proposed and applied to optimalthe fuzzy controller The structure of the paper is as follows inSection 2 the configuration of a series hybrid electric rescuevehicle is introduced in Section 3 the EMS of the wholevehicle is studied in Section 4 the method for optimizingfuzzy control based on IQGA is designed in Section 5 theoffline simulation andHILS experiment are described finallythe paper is concluded in Section 6

2 Configuration of the Series HybridElectric Rescue Vehicle

In series hybrid electric system there is no mechanicalconnection between the engine and the driving wheels thespeed and torque of the engine can be controlled to anyoperating point on its speed-torque map So by optimizingcontrol the fuel consumption and emission of the enginecan be improved As shown in Figure 1 the four-wheel drivestructure with two motors is adopted to the rescue vehicle inthis paper which can not only increase the vehicle power per-formance but also provide better regenerative braking effectThe electric drive system mainly includes general controllerinternal-combustion-engine-generator power unit (IGPU)and its controller battery pack and its management system(BMS) motors and motor controllers rectifier and DCDCconverter The general controller is the core componentwhich is used to coordinate the power distribution of IGPUand battery pack according to the driver instruction vehicle









Engine speed (rmin)

Gen

erat

or p

ower

(KW

)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259266

273

273

280308

800 1000 1200 1400 1600 1800 2000

50

100

150

200

250

300

350



ChargeDischarge

01 02 03 04 05 06 07 08 09 10SOC

001

002

003

004

005

006

007

008

009

01Re

sista

nce(

Ω)





Start

Pure electric

Electric mode A

Electric mode B



Hybrid mode A

Hybrid mode B

Braking mode A

Braking mode B

Return

Yes

No

Yes

Yes

No

No

NoYes

No

Yes

No

Yes

YesNo

Calculate Preq

SOC gt３ＦＩＱ

SOC gt３ＦＩＱ

Preq lt0

SOCge ３ＦＩＱ


Output Pg P

b PR

Preqgt50KW

①

⑧

⑦

④ ⑥

⑤

②

③



1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 0 (1)


1198751015840119887 = 119875119888max

1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (2)




1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (4)


1198751015840119887 = (1 minus 119896119892) sdot 1198751199031198901199021198751015840119892 = 119896119892 sdot 119875119903119890119902 (5)




VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

1203 04 05 06 07 08 09 1 110201

0

05

1

Deg

ree o

f mem

bers

hip

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

08 09 1 11 12 13 14 1507Kg






1198651205832 = 12 [[

119866ℎ119892radic1198872 + 4ℎ119892119871119866 1198651205831 minus (119866119887ℎ119892 + 21198651205831)]]

(6)



120573 sdot 119865119887119903119886 max = 1198651205831 + 1198651205832 (7)


01 02 03 04 05 06 07 08 100

01

02

03

04

05

06

09

a

b

I curve


j=03gj=02g

j=01g

Rear

bra

king

forc

e rat

io(F

2G)

Ffr2

FmＧＲ

FmＧＲ

Freg1

Freg2







1198961 = 1 0 le 119885 lt 070 119885 ge 07 (9)




1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)


1198963 =


(11)




119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)



1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)




1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)





1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)


119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512


100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)







follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018
















Engine speed (rmin)

Gen

erat

or p

ower

(KW

)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259266

273

273

280308

800 1000 1200 1400 1600 1800 2000

50

100

150

200

250

300

350



ChargeDischarge

01 02 03 04 05 06 07 08 09 10SOC

001

002

003

004

005

006

007

008

009

01Re

sista

nce(

Ω)





Start

Pure electric

Electric mode A

Electric mode B



Hybrid mode A

Hybrid mode B

Braking mode A

Braking mode B

Return

Yes

No

Yes

Yes

No

No

NoYes

No

Yes

No

Yes

YesNo

Calculate Preq

SOC gt３ＦＩＱ

SOC gt３ＦＩＱ

Preq lt0

SOCge ３ＦＩＱ


Output Pg P

b PR

Preqgt50KW

①

⑧

⑦

④ ⑥

⑤

②

③



1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 0 (1)


1198751015840119887 = 119875119888max

1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (2)




1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (4)


1198751015840119887 = (1 minus 119896119892) sdot 1198751199031198901199021198751015840119892 = 119896119892 sdot 119875119903119890119902 (5)




VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

1203 04 05 06 07 08 09 1 110201

0

05

1

Deg

ree o

f mem

bers

hip

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

08 09 1 11 12 13 14 1507Kg






1198651205832 = 12 [[

119866ℎ119892radic1198872 + 4ℎ119892119871119866 1198651205831 minus (119866119887ℎ119892 + 21198651205831)]]

(6)



120573 sdot 119865119887119903119886 max = 1198651205831 + 1198651205832 (7)


01 02 03 04 05 06 07 08 100

01

02

03

04

05

06

09

a

b

I curve


j=03gj=02g

j=01g

Rear

bra

king

forc

e rat

io(F

2G)

Ffr2

FmＧＲ

FmＧＲ

Freg1

Freg2







1198961 = 1 0 le 119885 lt 070 119885 ge 07 (9)




1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)


1198963 =


(11)




119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)



1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)




1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)





1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)


119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512


100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)







follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









Start

Pure electric

Electric mode A

Electric mode B



Hybrid mode A

Hybrid mode B

Braking mode A

Braking mode B

Return

Yes

No

Yes

Yes

No

No

NoYes

No

Yes

No

Yes

YesNo

Calculate Preq

SOC gt３ＦＩＱ

SOC gt３ＦＩＱ

Preq lt0

SOCge ３ＦＩＱ


Output Pg P

b PR

Preqgt50KW

①

⑧

⑦

④ ⑥

⑤

②

③



1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 0 (1)


1198751015840119887 = 119875119888max

1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (2)




1198751015840119887 = min (119875119903119890119902 119875119889max)1198751015840119892 = 119875119903119890119902 minus 1198751015840119887 (4)


1198751015840119887 = (1 minus 119896119892) sdot 1198751199031198901199021198751015840119892 = 119896119892 sdot 119875119903119890119902 (5)




VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

1203 04 05 06 07 08 09 1 110201

0

05

1

Deg

ree o

f mem

bers

hip

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

08 09 1 11 12 13 14 1507Kg






1198651205832 = 12 [[

119866ℎ119892radic1198872 + 4ℎ119892119871119866 1198651205831 minus (119866119887ℎ119892 + 21198651205831)]]

(6)



120573 sdot 119865119887119903119886 max = 1198651205831 + 1198651205832 (7)


01 02 03 04 05 06 07 08 100

01

02

03

04

05

06

09

a

b

I curve


j=03gj=02g

j=01g

Rear

bra

king

forc

e rat

io(F

2G)

Ffr2

FmＧＲ

FmＧＲ

Freg1

Freg2







1198961 = 1 0 le 119885 lt 070 119885 ge 07 (9)




1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)


1198963 =


(11)




119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)



1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)




1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)





1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)


119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512


100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)







follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

1203 04 05 06 07 08 09 1 110201

0

05

1

Deg

ree o

f mem

bers

hip

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

08 09 1 11 12 13 14 1507Kg






1198651205832 = 12 [[

119866ℎ119892radic1198872 + 4ℎ119892119871119866 1198651205831 minus (119866119887ℎ119892 + 21198651205831)]]

(6)



120573 sdot 119865119887119903119886 max = 1198651205831 + 1198651205832 (7)


01 02 03 04 05 06 07 08 100

01

02

03

04

05

06

09

a

b

I curve


j=03gj=02g

j=01g

Rear

bra

king

forc

e rat

io(F

2G)

Ffr2

FmＧＲ

FmＧＲ

Freg1

Freg2







1198961 = 1 0 le 119885 lt 070 119885 ge 07 (9)




1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)


1198963 =


(11)




119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)



1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)




1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)





1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)


119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512


100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)







follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018










1198962 =

1 V gt 24112119881 minus 1 12 lt 119881 le 24

0(10)


1198963 =


(11)




119875119903119890119902 = (11986510158401199031198901198921 + 11986510158401199031198901198922) sdot 1199031198940 sdot 119873 (13)



1198751015840119892 = minus119870119887119903119886119875119903119890119902 minus 11987510158401198871198751015840119877 = 0

(14)




1198751015840119892 = 01198751015840119877 = 119870119887119903119886119875119903119890119902 minus 1198751015840119887

(15)





1003816100381610038161003816120593⟩ = 120572 |0⟩ + 120573 |1⟩ (16)


119902119905119895 = [ 120572119905111205731199051110038161003816100381610038161003816100381610038161003816100381610038161205721199051212057311990512


100381610038161003816100381610038161003816100381610038161003816100381612057211990511198961205731199051119896

10038161003816100381610038161003816100381610038161003816100381610038161205721199052112057311990521

10038161003816100381610038161003816100381610038161003816100381610038161205721199052212057311990522


100381610038161003816100381610038161003816100381610038161003816100381612057211990521198961205731199052119896


100381610038161003816100381610038161003816100381610038161003816100381612057211990511989811205731199051198981

100381610038161003816100381610038161003816100381610038161003816100381612057211990511989821205731199051198982


1003816100381610038161003816100381610038161003816100381610038161003816120572119905119898119896120573119905119898119896 ] (17)







follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018












follows

119883119898 = 119880min119898 + ( 119896sum

119894=1


1198982119896 minus 1 (18)






120579 = Δ120579119894 sdot 119891 (120572119894 120573119894) (21)


120575 = 120579min




119891119886V119892 = sum119899119894=1 119891119894119899 (23)


Δ119891 = 10038161003816100381610038161003816119891119886V119892 minus 1198911015840119886V11989210038161003816100381610038161003816 (24)





VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









VS MS M MB VB

A E C

B

D

01+x1 01+x1+x2 01+x1+x2+x3 1201

002040608

1

Deg

ree o

f mem

bers

hip





119874119887119895119865119906119899 = 1055119862119894119905119910 119881119887 + 045119867119908119910 119881119887 (26)



119887119890119904119905and its fitness 1198741198871198951198651199061198991199050




Start




t=t+1

YES

NO




Obtain qt0best

ObjFunt0beat

Obtain ObjFunti












119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018














119879119908= [119898119892119862119903 cos 120572 + 121198621198631205881198601199062 + 119898119892 sin 120572 + 1205751015840119898119889119906119889119905 ] (28)


1198731 = 1198732 = 301198940119906120587119903 (29)


+

-

+

minus

kx g

ke g

g

Tg

Idc

Udc


119880119889119888 = 119870119890120596119892 minus 119870119909120596119892119868119889119888119879119892 = 1198751015840119892120596119892 = 119870119890119868119889119888 minus 1198701198901198682119889119888 (30)


119879119890119899119892119894119890119892 minus 119879119892 = 01047 ( 1198691198901198942119890119892 + 119869119892) 119889119899119892119889119905119899119892 = 119894119890119892119899119890119899119892

(31)



+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









+

-

+

minus

Rdcℎ

Rcℎ

V0

Ub

Ib



119881119890 = 1198871198901198791198901198991198921198991198901198991198929550120578119890120588119890 (32)




119877119889119888ℎ = 1198912 (SOC)119877119888ℎ = 1198913 (SOC)

(33)


119878119874119862 = 1198760 minus int1199050

119868119887 (119905) 119889119905119876max

times 100 (34)


500 1000 1500 2000 25000Time (s)

0

20

40

60

80

100

Velo

city

(km

h)


20 40 60 80 100 120 140 160 180 2000Iteration times

35

4

45

5

55

6

Glo

bal b

est fi

tnes

s

GAQGAIQGA





VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









VS MS M MB VB

VS MS M MB VB

VS MS M MB VB

08 09 1 11 12 13 14 1507Kg

0

05

1

Deg

ree o

f mem

bers

hip

055 06 065 07 075 0805SOC

0

05

1

Deg

ree o

f mem

bers

hip

02 03 04 05 06 07 08 09 1 11 1201

0

05

1

Deg

ree o

f mem

bers

hip


Engine speed (rmin)

199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280308

287

800 1000 1200 1400 1600 1800 2000

100

150

50

200

250

300

350

Gen

erat

or p

ower

(KW

)







50

100

150

200

250

300

350

Gen

erat

or p

ower

(KW

)


199

200

201

203

205

206

208

210

217

224

231

238

245

252

252

259

259

266

266

273

273

280287

308


500 1000 1500 2000 25000Time (s)

067

068

069

07

071

072

SOC



0123456789

Fuel

cons

umpt

ion６

(L

)

500 1000 1500 2000 25000Time (s)




Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









Driver

AD converter

Control strategy

CAN bus transceiver

CAN bus transceiver


CAN bus transceiver

Monitor

dSPACEMicroAutoBox

CAN bus

CAN bus

Host computer


Operating signal

General controller

BatteryMotor2


dSPACEMicroAutoBox

DC power

Generalcontroller

Driver input

Monitor

Hostcomputer









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018









10 20 30 40 50 60 700Time (s)

minus15

minus1

minus05

0

05

1

15Pe

dal p

ositi

on

(a)

0

20

40

60

80

100

Velo

city

(km

h)

10 20 30 40 50 60 700Time (s)

(b)

Preq

Pg

Pb

10 20 30 40 50 60 700Time (s)

minus400

minus200

0

200

400

Pow

er (K

W)

(c)


minus15000

minus10000

minus5000

0

5000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(d)


minus8000

minus6000

minus4000

minus2000

0

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(e)


minus4000

minus3000

minus2000

minus1000

0

1000

2000

Torq

ue (N

m)

10 20 30 40 50 60 700Time (s)

(f)



6 Conclusions




Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018










Data Availability




Acknowledgments


References




















































Journal of












Journal of







Volume 2018






Volume 2018



































Journal of












Journal of







Volume 2018






Volume 2018








a novel energy management strategy for series hybrid...

Documents