Electrical Power and Energy Systems 43 (2012) 514–525

Contents lists available at SciVerse ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

Energy management strategy for fuel cell/battery/ultracapacitor hybrid vehicle

based on fuzzy logic
Qi Li a,⇑, Weirong Chen a, Yankun Li a, Shukui Liu b, Jin Huang a
a
School of Electrical Engineering, Southwest Jiaotong University, Chengdu 610031, Sichuan Province, China
b
Chengdu Electric Power Bureau, Chengdu 610021, Sichuan Province, China

a r t i c l e i n f o a b s t r a c t

Article history: In order to enhance the fuel economy of hybrid vehicle and increase the mileage of continuation of jour-
Received 31 August 2010 ney, a fuzzy logic control is utilized to design energy management strategies for fuel cell/battery (FC + B)
Received in revised form 20 May 2012 hybrid vehicle and fuel cell/battery/ultra-capacitor (FC + B + UC) hybrid vehicle. The models of hybrid
Accepted 2 June 2012
vehicle for FC + B and FC + B+UC structure are developed by electric vehicle simulation software ADVISOR
Available online 7 July 2012
which uses a hybrid backward/forward approach. The results demonstrate that the proposed control
strategy can satisfy the power requirement for four standard driving cycles and achieve the power dis-
Keywords:
tribution among various power sources. The comprehensive comparisons with the power tracking con-
Hybrid vehicle
Energy management
trol strategy which is wide adopted in ADVISOR verify that the proposed control strategy has better
Fuzzy logic rationality and validity in terms of fuel economy and dynamic property in four standard driving cycles.
Fuel cell Therefore, the proposed strategy will provide a novel approach for the advanced energy management sys-
Battery tem of hybrid vehicle.
Ultra-capacitor Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction dynamic characteristic of fuel cell, a fuel cell/battery (FC + B) or

fuel cell/ultra-capacitor (FC + UC) hybrid vehicle has been pro-
Hybrid vehicles are vehicles that use two or more power posed [5,22]. The energy management of hybrid vehicle which
sources for the drive system [1]. In contrast, ordinary internal com- determines the power assignment between the fuel cell system
bustion engine (ICE) vehicles use a single power source consisting and auxiliary energy storage devices is an important technique.
of reciprocating engine, typically fueled with gasoline, to drive a In recent years, a variety of control strategies for energy man-
complex transmission mechanism that is then coupled to the drive agement have been used to hybrid vehicle. Kim et al. [5] had pro-
wheels. The disadvantages of ICE vehicles include low energy effi- posed a fuzzy controller to optimally distribute the relative power
ciency, excessive harmful chemical emissions, high noise level and between the fuel cell and the battery. Thounthong et al. [6] had
heavy dependence on a single fuel source. Hybrid electric vehicles used an innovative control law based on flatness properties for fuel
are one of the solutions proposed to tackle the perceived problems cell/supercapacitor hybrid power source. Paladini et al. [7] had pre-
associated with the energy crisis and global warming. Hybrid vehi- sented an optimal control strategy to power a vehicle with both
cles seamlessly combine two or more power sources into one drive fuel cell and battery to reduce fuel consumption. Schiffer et al.
system [2–4]. Fuel cell hybrid vehicles merge hybrid vehicle and [8] had reported a state control strategy for FC + UC hybrid vehicle
the hydrogen fuel cell technologies in order to replace the conven- based on different driving phases to minimize the hydrogen con-
tional fuel and avoid CO2 emissions and optimize the fuel sumption and to assure power availability at any time. However,
consumption [5–9]. in these works the proposed control strategies had not adequately
As a promising technology fuel cells that convert chemical en- considered the balance between fuel economy and dynamic prop-
ergy of the fuel into electricity without combustion are studied erty of hybrid vehicle [5–10]. Furthermore, an appropriate intelli-
worldwide with an aim to improve the power output, lower the gent control strategy had not been proposed for a hybrid vehicle.
cost and extend the life of operation for widespread applications Therefore, in order to enhance the fuel economy and increase the
[10–14]. The fuel cells are generally viewed as a dependable power mileage of continuation of journey for FC + B+UC hybrid vehicle,
source for many applications, such as hybrid vehicle, distributed a fuzzy logic control method is utilized to design relevant energy
power generation, and portable power source [15–21]. Due to slow management strategy.
One of the main objectives of the energy management strategy
⇑ Corresponding author. Tel.: +86 2887603332; fax: +86 2887605114. is to manage power sharing between various components and to
E-mail address: [email protected] (Q. Li). guarantee the power sources performance. Fuzzy logic control is

0142-0615/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijepes.2012.06.026
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 515

basically an adaptive and nonlinear control which gives robust per-

formance for linear or nonlinear plants [22]. The advantages that
this control strategy has are listed as follows [5]: the fuzzy logic
basis is easy to understand and is expressed in natural language;
fuzzy logic is flexible and can deal with imprecise data; fuzzy logic
does not necessitate a mathematical model of the system to be
controlled; fuzzy logic can be used as a tool to incorporate experi-
ence of expert in the control strategy; fuzzy logic can be imple-
mented together with conventional control techniques without
any kind of incompatibility. As the hybrid system is nonlinear
and multivariable, the fuzzy logic control strategy is more suitable
for the energy management.
In this paper, a secondary development for electric vehicle sim-
ulation software ADVISOR is implemented based on the system
architecture of FC + B hybrid vehicle. The loading files are reconfig-
uration and then a model of FC + B + UC hybrid vehicle is devel-
oped. In order to enhance the fuel economy of hybrid vehicle and
increase the mileage of continuation of journey, a fuzzy logic con-
trol method is utilized to design relevant energy management
strategy for FC + B hybrid vehicle and FC + B + UC hybrid vehicle. Fig. 2. FC + B + UC drive structure of hybrid vehicle.
According to standard driving cycle conditions, the proposed con-
trol strategy is contrasted with the power tracking control strategy and ultra-capacitor jointly provide supplemental power, such as
which is wide adopted in ADVISOR in terms of the indexes of fuel startup, acceleration, and climbing of vehicle. The ultra-capacitor
economy and dynamic property. parallelly connects the DC bus by a bidirectional DC/DC converter.
The advantages of this structure are that the ultra-capacitor can
provide the peak power and recover braking energy so that the
2. Drive structure of hybrid vehicle burden of fuel cell system and battery are alleviated, and also pro-
long the working life of battery by less charging and discharging.
2.1. Fuel cell/battery (FC + B) structure However, the structure and control strategy are complex.

With the advancement in the technology of fuel cells, there is an

3. Energy management strategy for hybrid vehicle
increasing interest in using fuel cells for hybrid vehicle. A FC + B
drive structure is a popular hybrid structure as Fig. 1 shown [7–
As the hybrid vehicle is nonlinear and multivariable, the fuzzy
9]. In this structure, a fuel cell system designed for vehicular pro-
logic controller (FLC) is more suitable for the energy management.
pulsion application must have a power density, a startup, and a
The controller relates its output to inputs using a list of IF-THEN
transient response similar to present-day ICE-based vehicles. A
rules. The IF part of a rule, specifies the condition for which a rule
battery is generally connected across the fuel cell system to pro-
holds. The THEN part of a rule refers to values of the output vari-
vide supplemental power for starting the system. The fuel cell sys-
able to obtain the output of the controller. A degree of membership
tem and inverter connect by a unidirectional DC/DC converter for
is assigned to the variables according to the membership functions
matching voltage class. The advantages of this structure are low
definition. The membership degree of the IF part of all rules are
power and transient response demand from the fuel cell system
evaluated and all rules of the THEN part are averaged and weighted
and convenient braking energy recovery. But the maintenance cost
by these membership degrees [5].
of battery is increased.

3.1. Fuzzy logic control for FC + B

2.2. Fuel cell/battery/ultra-capacitor (FC + B + UC) structure
In order to rationally assign the demand power of electrical mo-
According to the FC + B structure, a FC + B + UC drive structure tor to fuel cell system and battery, the demand power of electrical
has been proposed as Fig. 2 shown. In this structure, the battery motor and the state of charge (SOC) of battery are treated as two
input variables by fuzzy processing and the output variable is re-
quired power of fuel cell system. Four driving modes are adopted
for FC + B.

(1) Starting mode

In this mode, the battery individually drives the vehicle. The

fuel cell system should be activated according to the demand
power and SOC of battery. The balance relation of the gross power
is expressed as follows:
8 Pr
< P m ¼ gm
>
Pfc ¼ 0 ð1Þ
>
:
PB ¼ Pm
where Pr is the demand power of working condition, Pfc is the out
power of fuel cell system, Pm is the motor power, PB is the battery
Fig. 1. FC + B drive structure of hybrid vehicle. power, and gm is the motor efficiency.
516 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525

Fig. 3. Membership functions of input and output variables.

8 
(2) Fuel cell driving and battery charging mode < PB ¼ ðSOCSOC Þ Pc
0:5ðSOC SOC Þ
h l
ð3Þ
:
In this mode, the fuel cell system individually drives the vehicle. Pfc gDC1 ¼ P m  PB
If the SOC value is lower than the expected SOC value and the de-
Furthermore, if the demand power is greater than the maxi-
mand power of working condition locates in the range of maxi-
mum power of fuel cell system, the battery must discharge no mat-
mum and minimum out power of the fuel cell system, the out
ter SOC is or is not higher than SOC⁄. The fuel cell system provides
power of fuel cell system must satisfy the demand power and
the demand power with maximum power and the surplus demand
charge the battery. The charging power of battery depends on
power is provided by the battery. The balance relation is shown as
the difference of the present SOC and the expected SOC. The bal-
follows:
ance relation of the gross power is shown as follows:
8 ðSOCSOC  Þ (
< PB ¼ 0:5ðSOC h SOC l Þ Pc
> Pfc ¼ Pfc;max
ð2Þ ð4Þ
> Pfc gDC1 ¼ Pm þ PB PB ¼ Pm  Pfc gDC1
:
SOC  ¼ 0:5ðSOC h þ SOC l Þ
where SOC⁄ is the expected SOC value, SOCh and SOCl are the set va- (4) Regenerative braking mode
lue of maximum and minimum, Pc is the regulation power of SOC,
and gDC1 is the efficiency of unidirectional DC/DC converter. In this mode, the motor turns into power generation mode as
the vehicle decelerates or brakes. The battery recovers energy
(3) Fuel cell and battery combined driving mode and the fuel cell system shuts down.
The FLC for FC + B has two input variables and one output var-
In this mode, if SOC is higher than SOC⁄, the battery must dis- iable. The fuzzy field scope of the motor demand power Pm is set as
charge and the surplus demand power is provided by the fuel cell [1, 1], the fuzzy field scope of SOC is set as [0, 1], and the fuzzy
system. The discharging power of battery depends on the differ- field scope of output power of fuel cell Pfc is set as [0, 1]. The fuzzy
ence of the present SOC and SOC⁄. The balance relation of the gross subset of Pm is divided into {NH, NM, NL, PL, PM, PH}, the fuzzy
power is expressed as follows: subset of SOC is divided into {L, M, H}, the fuzzy subset of Pfc is di-
vided into {L0, L1, L2, L3, L4, L5}. In this paper, the membership
functions are achieved by inhomogeneous distribution to avoid
over charging and discharging of battery as Fig. 3 shown. In
Fig. 3, the Trapezoid Membership Function is employed for Pm,
the S-type Membership Function, the Z-type Membership Function
and the Triangle Membership Function are employed together for
SOC, the Trapezoid Membership Function is also employed for Pfc.
Fig. 4 is the surface chart of output characteristic.
The fuzzy reasoning rules with 18 items are established for
manipulated variables as Table 1 shown. The form of fuzzy reason-
ing is ‘‘IF Pm is Ai AND SOC is Bi, Then Pfc is Ci’’. All rules are on the

Table 1
Fuzzy reasoning rules.

Pm
SOC NH NM NL PL PM PH
L L0 L1 L2 L3 L4 L5
M L0 L0 L1 L2 L3 L5
H L0 L0 L0 L1 L3 L4
Fig. 4. Surface chart of output characteristic.
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 517

Fig. 5. Membership functions of input and output variables for No. 1 FLC.

Fig. 6. Membership functions of input and output variables for No. 2 FLC.

basis of practical experience. Mamdani Inference Method is uti- (1) Starting mode
lized to carry out the fuzzy reasoning.
The method of weighted mean (centroid method) is used to In this mode, the battery and ultra-capacitor combined drive
convert the vector into a single value and then the defuzzification the vehicle. The fuel cell system should be activated according to
is achieved. the demand power and SOC of battery and ultra-capacitor after
R preheating. The balance relation of the gross power is expressed
xi  lðxi Þdxi as follows:
u0 ¼ R ð5Þ
lðxi Þdxi 
Pfc ¼ 0
ð6Þ
where xi is the element of a set in the field, l(xi) is the membership PB þ P UC gDC2 ¼ Pm
degree, u0 is the judgment result of fuzzy output.
where PUC is the ultra-capacitor power, and gm is the efficiency of
bidirectional DC/DC converter.
3.2. Fuzzy logic control for FC + B + UC

(2) Fuel cell individual driving mode

In order to rationally assign the demand power of electrical mo-
tor to fuel cell, battery and ultra-capacitor for FC + B + UC, two
In this mode, the fuel cell system individually drives the vehicle.
decentralized controllers are designed according to five driving
If the SOC value of battery and the CSOC value of ultra-capacitor
modes as following shown:

Table 3
Table 2
Part fuzzy reasoning rules of No. 2 FLC.
Part fuzzy reasoning rules of No. 1 FLC.
1. If (Ps is NH) and (SOC is H) and (CSOC is H) then (PB is Z)
1. If (Pm is NH) and (SOC is H) and (CSOC is H) then (Pfc is L0)
2. If (Ps is NM) and (SOC is M) and (CSOC is M) then (PB is NL)
2. If (Pm is NM) and (SOC is M) and (CSOC is M) then (Pfc is L1)
3. If (Ps is NL) and (SOC is L) and (CSOC is L) then (PB is Z)
3. If (Pm is NL) and (SOC is L) and (CSOC is L) then (Pfc is L2)
4. If (Ps is PL) and (SOC is L) and (CSOC is L) then (PB is PL)
4. If (Pm is PL) and (SOC is L) and (CSOC is L) then (Pfc is L2)
5. If (Ps is PM) and (SOC is M) and (CSOC is M) then (PB is PM)
5. If (Pm is PM) and (SOC is M) and (CSOC is M) then (Pfc is L3)
6. If (Ps is PH) and (SOC is H) and (CSOC is H) then (PB is PM)
6. If (Pm is PH) and (SOC is H) and (CSOC is H) then (Pfc is L3)
518 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525

Fig. 7. ADVISOR vehicle input screen.

Fig. 8. Simulation model of FC + B + UC.

are lower than SOC⁄ value and the demand power of working con- satisfy the demand power and charge the battery and ultra-capac-
dition locates in the range of maximum and minimum out power itor. The prior charging power must deliver to ultra-capacitor so
of the fuel cell system, the out power of fuel cell system must that the demand peak power of electrical motor could be provided
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 519

Fig. 9. Simulink model of fuel cell system.

Fig. 10. Simulink model of battery.

quickly. And then the battery can be charged while CSOC reachs to (4) Fuel cell, battery and ultra-capacitor combined driving mode
SOC⁄. The balance relation of the gross power is shown as follows:
In this mode, if the demand power is greater than the maximum
Pfc gDC1 ¼ Pm þ PB þ PUC gDC2 ð7Þ
power of fuel cell system, the fuel cell system provides the demand
power with maximum power and the surplus demand power is
provided by the ultra-capacitor. If the demand power still cannot
(3) Fuel cell and battery combined driving mode
be satisfied, the battery must discharge whether SOC is or is not
higher than SOC⁄. The balance relation is shown as follows:
In this mode, if SOC of battery is higher than SOC⁄, the battery
must discharge and the surplus demand power is provided by 
Pfc ¼ Pfc;max
the fuel cell system. If CSOC of ultra-capacitor is lower than SOC⁄, ð9Þ
Pfc gDC1 þ PB þ PUC gDC2 ¼ Pm
the ultra-capacitor must charge by the fuel cell system. The bal-
ance relation is expressed as follows:
Pfc gDC1 þ PB ¼ Pm þ PUC gDC2 ð8Þ (4) Regenerative braking mode
520 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525

Fig. 11. Simulink model of ultra-capacitor.

In this mode, the motor turns into power generation mode as fuzzy field scope of CSOC is set as [0, 1], and the fuzzy field scope of
the vehicle decelerates or brakes. The ultra-capacitor recovers en- output power of fuel cell Pfc is set as [0, 1]. The fuzzy subset of Pm is
ergy firstly. The battery recovers energy as the ultra-capacitor is divided into {NH, NM, NL, PL, PM, PH}, the fuzzy subsets of SOC and
full of energy and the fuel cell system shuts down. CSOC are divided into {L, M, H}, the fuzzy subset of Pfc is divided
The No.1 FLC for FC + B + UC has three input variables and one into {L0, L1, L2, L3, L4, L5}. For No.2 FLC, the fuzzy field scope of
output variable. The input variables include the demand power surplus demand power Ps is set as [1, 1], the fuzzy field scope
of electrical motor, SOC of battery and CSOC of ultra-capacitor. of SOC is set as [0, 1], the fuzzy field scope of CSOC is set as [0,
The output variable is the required power of fuel cell system. This 1], and the fuzzy field scope of output power of battery PB is set
FLC can rationally assign the demand power of motor to fuel cell as [1, 1]. The fuzzy subset of Ps is divided into {NH, NM, NL, PL,
system and other supplementary power sources according to cur- PM, PH}, the fuzzy subsets of SOC and CSOC are divided into {L,
rent demand power of motor, SOC and CSOC. The No. 2 FLC for M, H}, the fuzzy subset of PB is divided into {NH, NM, NL, Z, PL,
FC + B + UC has three input variables and one output variable. PM, PH}. The membership functions are achieved by inhomoge-
The input variables include surplus demand power of motor neous distribution as Figs. 5 and 6 shown. In Figs. 5 and 6, the Trap-
(Ps = Pm  Pfc), SOC and CSOC. The output variable is the output ezoid Membership Functions are employed for Pm and Ps, the S-
power of battery. This FLC can rationally assign surplus demand type Membership Function, the Z-type Membership Function and
power of motor to battery and ultra-capacitor according to current the Triangle Membership Function are employed together for
surplus demand power of motor, SOC and CSOC. The control objec-
tive with the two FLC would be achieved by reducing hydrogen
Table 5
consumption on condition that dynamic property of vehicle is Main hybrid vehicle parameters.
satisfied.
Vehicle
For No. 1 FLC, the fuzzy field scope of the motor demand power
Mass (kg) 1191 Coefficient of rolling drag 0.009
Pm is set as [1, 1], the fuzzy field scope of SOC is set as [0, 1], the Frontal area (m2) 2.0 Coefficient of aerodynamic 0.335
drag
Wheel rolling radius (m) 0.282 Wheelbase (m) 2.6

Table 4 Motor
Parameters of driving cycle condition. Maximum power (kW) 75 Maximum speed (rpm) 6283
Rated voltage (V) 320 Average efficiency (%) 90
Parameters UDDS HWFET US06 ECE + EUDC
Fuel cell system
Time (s) 1369 765 600 1225 Type PEM fuel cell
Distance (km) 11.99 16.51 12.89 10.93 Maximum net power (kW) 50 Average efficiency (%) 56
Average apeed (km/h) 31.51 77.57 77.20 32.11
Battery
Maximum speed (km/h) 91.25 96.40 129.23 120.00
Maximum discharging 5C Rated voltage (V) 12
Average acceleration (m/s2) 0.51 0.20 0.67 0.54
rate
Average deceleration (m/s2) 0.58 0.22 0.73 0.79
Capacity 25Ah Number 25
Maximum acceleration (m/s) 1.48 1.43 3.76 1.05
Maximum deceleration (m/s2) 1.48 1.48 3.08 1.39 Ultra-capacitor
Idle time (s) 259 6 45 339 Type PC2500
Number of stop 17 1 5 13 Capacity 2500F
Grade (%) 0 0 0 0 Maximum voltage (V) 2.5 Number 155
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 521

(a) UDDS (b) HWFET

(c) US06 (d) ECE+EUDC

Fig. 12. Comparisons between required speed and actual speed.

Table 6
Test results of fuel economy.

Cycle condition Economy (L/100 km) FC + B (PTC) FC + B (FLC) FC + B + UC (FLC)

UDDS Hydrogen consumption 75.9 71.2 49.0
Gasoline equivalent 5.1 4.8 3.3
HWFET Hydrogen consumption 42.0 39.2 30.9
Gasoline equivalent 2.9 2.6 2.1
US06 Hydrogen consumption 54.7 48.0 45.2
Gasoline equivalent 3.7 3.3 3.0
ECE + EUDC Hydrogen consumption 71.3 65.3 44.3
Gasoline equivalent 4.9 4.4 2.9

SOC and CSOC, the Trapezoid Membership Functions are also em- 4. Modeling of hybrid vehicle
ployed for Pfc and PB.
In order to achieve the control objective, alleviate the burden of 4.1. Principle of ADVISOR
fuel cell system and prolong working life of battery, the fuzzy rea-
soning rules with 54 items are respectively established for two FLC. The simulation package Advanced Vehicle Simulator (ADVISOR)
The part rules are listed in Tables 2 and 3. Similarly, Mamdani was originally developed by the National Renewable Energy Labo-
Inference Method is also utilized to carry out the fuzzy reasoning ratory (NREL) in 1994 to meet the growing need of the transporta-
and then the method of weighted mean (centroid method) is used tion industry for powertrain simulation [22]. ADVISOR is a Matlab/
to implement the defuzzification. Simulink program that relies on a central Simulink model for ac-
522 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525

(a) UDDS

(b) HWFET
Fig. 13. Power curves of FC + B + UC hybrid vehicle.

tual simulation calculations. ADVISOR uses a hybrid backward/for- circuit model based on PC2500 model of Maxwell Company is uti-
ward approach that is closely related to the strictly backward-fac- lized to develop the ultra-capacitor model as Fig. 11 shown. The
ing approach. The collections of ancillary Matlab scripts load all the validity of these models was verified by plentiful experiments [22].
necessary data and setup all of the inputs for the Simulink model.
Fig. 7 shows the vehicle input screen from ADVISOR. This screen is
used to load all of the configuration data and start the simulation in 4.2. Simulation parameters
the background.
In this paper, a secondary development for ADVISOR is imple- In this paper, Urban Dynamometer Driving Schedule (UDDS),
mented based on the system architecture of FC + B hybrid vehicle. Highway Fuel Economy Test (HWFET), Urban Schedule 06 (US06),
The loading files are reconfiguration and then a model of Economic Commission for Europe and Extra Urban Driving Cycle
FC + B + UC hybrid vehicle is developed as Fig. 8 shown. A fuel cell (ECE + EUDC) are selected as four standard cycle conditions for
system model based on 50 kW power-efficiency model of Interna- testing vehicle performance. The cycle parameters are shown in
tional Fuel Cell Company is adopted as Fig. 9 shown. A resistance– Table 4. According to the suggestions of Refs. [5,7], the default va-
capacitance (RC) equivalent circuit model is used to develop the lue by ADVISOR is utilized to determine the vehicle parameters.
battery model as Fig. 10 shown. An internal resistance equivalent Therefore, the degree of hybridization is 51.6%, initial values of
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 523

(c) US06

(d) ECE+EUDC
Fig. 13 (continued)
Table 7
Test results of dynamic property.

SOC and CSOC are 0.7, SOCh and SOCl are 0.8 and 0.4. The main hy- Dynamic property FC + B FC + B FC + B + UC
brid vehicle parameters are shown in Table 5. (PTC) (FLC) (FLC)
0–100 km/h Acceleration time 10.4 9.8 9.9
(s)
5. Results and discussion 60–100 km/h Acceleration time 5.3 5.2 5.9
(s)
0–140 km/h Acceleration time 20.5 22.5 17.3
In order to compare the indexes of fuel economy and dynamic (s)
property of hybrid vehicle, the FLCs are designed for FC + B hybrid Maximum speed (km/h) 156.9 157.4 155.5
vehicle and FC + B + UC hybrid vehicle. In addition, according to Maximum acceleration (m/s2) 4.9 4.9 4.9
Distance in 5 s (m) 52.9 55.2 58.7
four standard cycle conditions, the proposed control strategies
400 m Acceleration time (s) 17.6 17.4 16.9
are contrasted with the power following control strategy which Gradeability (%) 50.1 50.1 50.1
is wide adopted in ADVISOR for FC + B. The speed curves of FLC
524 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525

(a) UDDS (b) HWFET

(c) US06 (d) ECE+EUDC

Fig. 14. Comparisons of hydrogen consumption.

for FC + B, FLC for FC + B + UC, and power tracking controller (PTC) are alleviated. In regenerative braking condition, the ultra-capaci-
for FC + B can match with the required speed curves in four cycle tor recovers energy firstly and the battery recovers energy while
conditions. Therefore, three controllers designed can satisfy the the ultra-capacitor is full of energy. Therefore, the ultra-capacitor
speed requirements for four standard cycle conditions. The speed in FC + B + UC structure can give full play to the characteristic of
curves are shown in Fig. 12. fast charging and discharging and prolong the working life of bat-
In Table 6, the boldtypes in the columns represent the optimal tery by less charging and discharging. Comparisons of hydrogen
value of fuel economy. In different cycle conditions, the FLC for consumption in four cycle conditions are shown in Fig. 14. The
FC + B has smaller consumption (L/100 km) than the PTC for FLC for FC + B + UC has the smallest hydrogen consumption under
FC + B. The FLC for FC + B + UC has the smallest consumption (L/ all driving cycles. Therefore, this method can achieve better fuel
100 km) than other methods. Therefore, the conclusion can be economy.
drawn that the FLC for FC + B + UC has better performance in terms
of fuel economy. In Table 7, the boldtypes represent the optimal 6. Conclusions
value of dynamic property. The FLC for FC + B + UC has better per-
formance than other methods in 0–140 km/h acceleration time, In this paper, a fuzzy logic control method which is imple-
distance in 5 s, and 400 m acceleration time. However, the FLC mented in ADVISOR environment is utilized to design relevant en-
for FC + B has better performance than other methods in 0– ergy management strategies for FC + B hybrid vehicle and
100 km/h acceleration time, 60–100 km/h acceleration time, and FC + B + UC hybrid vehicle for the improvement of fuel economy
maximum speed. Furthermore, the maximum acceleration and and mileage of continuation of journey. A secondary development
gradeability are same for three methods. for ADVISOR is implemented based on the system architecture of
Fig. 13 shows power curves of FC + B + UC hybrid vehicle. In FC + B hybrid vehicle. The loading files are reconfiguration and
Fig. 13, the ultra-capacitor can provide the peak power for acceler- then a model of FC + B + UC hybrid vehicle is developed. The results
ation condition so that the burden of fuel cell system and battery indicate that the proposed control strategy can satisfy the power
 Q. Li et al. / Electrical Power and Energy Systems 43 (2012) 514–525 525

requirement for four standard driving cycles. In four cycle condi- [6] Thounthong P, Pierfederici S, Martin J-P, Hinaje M, Davat B. Modeling and
control of fuel cell/supercapacitor hybrid source based on differential flatness-
tions, the FLC for FC + B has smaller consumption (L/100 km) than
control. IEEE Trans Veh Technol 2010;59(6):2700–10.
the PTC for FC + B. The FLC for FC + B + UC has the smallest con- [7] Paladini V, Donateo T, de Risi A, Laforgia D. Control strategy optimization of a
sumption (L/100 km) than other methods. Hence, the conclusion fuel-cell electric vehicle. J Fuel Cell Sci Technol 2008;5:12–9.
can be drawn that the FLC for FC + B + UC has better performance [8] Schiffer J, Bohlen O, De Doncker RW, Sauer DU, Kyun Young Ahn. Optimized
energy management for fuelcell-supercap hybrid electric vehicles. In: IEEE
in terms of fuel economy. Furthermore, the FLC for FC + B + UC conference on vehicle power and propulsion; 2005. p. 341–8.
has better performance in terms of dynamic property in four stan- [9] Markel T, Brooker A, Hendricks T. A systems analysis tool for advanced vehicle
dard driving cycle conditions. Hence, the proposed strategy will modeling. J Power Sour 2002;110:255–66.
[10] Emadi Ali, Lee Young Joo, Rajashekara Kaushik. Power electronics and motor
give a novel approach for the advanced energy management sys- drives in electric, hybrid electric, and plug-in hybrid electric vehicles. IEEE
tem of hybrid vehicle. Trans Indust Electron 2008;55(6):2237–45.
[11] Jin K, Ruan X, Yang M, Xu M. A hybrid fuel cell power system. IEEE Trans Indust
Electron 2009;56(4):1212–22.
Acknowledgment [12] Dai Chaohua, Chen Weirong, Cheng Zhanli, Li Qi, Jiang Zhiling, Jia Junbo. Seeker
optimization algorithm for global optimization: a case study on optimal
The authors thank Prof. Youyi Wang and the reviewers for their modelling of proton exchange membrane fuel cell (PEMFC). J Electr Power
Energy Syst 2011;33:369–76.
helpful suggestions. This work was supported by National Natural [13] Wang Yong, Choi Seeyoung, Lee Eunchul. Efficient and ripple-mitigating dc–dc
Science Foundation of China (51177138), Specialized Research Fund converter for residential fuel cell system. J Electr Power Energy Syst
for the Doctoral Program of Higher Education (20100184110015), 2009;32:43–9.
[14] Li Q, Chen W, Wang Y, Liu S, Jia J. Parameter identification for PEM fuel cell
the Fundamental Research Funds for the Central Universities
mechanism model based on effective informed adaptive particle swarm
(SWJTU11CX030) and International Science and Technology Coop- optimization. IEEE Trans Indust Electron 2011;58(6):2410–9.
eration and Exchange Research Plan of Sichuan Province [15] Li Q, Chen W, Wang Y, Jia J, Han M. Nonlinear robust control of proton
(2012HH0007). Science and Technology Development Plan of Min- exchange membrane fuel cell by state feedback exact linearization. J Power
Sour 2009;194(1):338–48.
istry of Railways ( 2012J012-D ). [16] Jia J, Li Q, Wang Y, Cham YT, Han M. Modeling and dynamic characteristic
simulation of proton exchange membrane fuel cell. IEEE Trans Energy Convers
References 2009;24(1):283–91.
[17] Zhao Hengbing, Burke Andrew F. Optimization of fuel cell system operating
conditions for fuel cell vehicles. J Power Sour 2009;186:408–16.
[1] Ramos-Paja CA, Romero A, Giral R, Calvente J, Martinez-Salamero L. [18] Palma L, Todorovic MH, Enjeti PN. Analysis of common-mode voltage in
Mathematical analysis of hybrid topologies efficiency for PEM fuel cell utility-interactive fuel cell power conditioners. IEEE Trans Indust Electron
power systems design. J Electr Power Energy Syst 2010;32:1049–61.
2009;56(1):20–7.
[2] Hajizadeh Amin, Golkar Masoud Aliakbar. Control of hybrid fuel cell/energy [19] Wang C, Nehrir MH, Gao H. Control of PEM fuel cell distributed generation
storage distributed generation system against voltage sag. J Electr Power
systems. IEEE Trans Energy Convers 2006;21(2):586–95.
Energy Syst 2010;32:488–97. [20] Na Woon Ki, Gou Bei. Feedback-linearization-based nonlinear control for PEM
[3] Hajizadeh Amin, Golkar Masoud Aliakbar. Intelligent power management fuel cells. IEEE Trans Energy Convers 2008;23(1):179–90.
strategy of hybrid distributed generation system. J Electr Power Energy Syst
[21] Suh Kyung-Won, Stefanopoulou AG. Performance limitations of air flow
2007;29:783–95. control in power-autonomous fuel cell systems. IEEE Trans Control Syst
[4] Dursun Erkan, Kilic Osman. Comparative evaluation of different power
Technol 2007;15(3):464–73.
management strategies of a stand-alone PV/Wind/PEMFC hybrid power [22] Wipke KB, Cuddy MR, Burch SD. ADVISOR 2.1: a user-friendly advanced
system. J Electr Power Energy Syst 2012;34:81–9. powertain simulation using a combined backward/forward approach. IEEE
[5] Kim M, Sohn YJ, Lee WY, Kim CS. Fuzzy control based engine sizing optimization Trans Veh Technol 1999;48(6):1751–61.
for a fuel cell/battery hybrid mini-bus. J Power Sour 2008;178:706–10.