Energy Conversion and Management: Chen LV, Junzhi Zhang, Yutong Li, Ye Yuan
Energy Conversion and Management: Chen LV, Junzhi Zhang, Yutong Li, Ye Yuan
Energy Conversion and Management: Chen LV, Junzhi Zhang, Yutong Li, Ye Yuan
a r t i c l e i n f o a b s t r a c t
Article history: This paper mainly focuses on control algorithm of the braking energy regeneration system of an electric
Received 5 July 2015 bus under safety–critical driving situations. With the aims of guaranteeing vehicle stability in various
Accepted 23 September 2015 types of tyre–road adhesion conditions, based on the characteristics of electrified powertrain, a novel
Available online 23 October 2015
control algorithm of regenerative braking system is proposed for electric vehicles during anti-lock brak-
ing procedures. First, the models of vehicle dynamics and main components including braking energy
Keywords: regenerative system of the case-study electric bus are built in MATLAB/Simulink. Then, based on the
Regenerative braking
phase-plane method, the optimal brake torque is calculated for ABS control of vehicle. Next, a novel
Electric vehicle
Safety–critical driving maneuvers
allocation strategy, wherein the target optimal brake torque is divided into two parts that are handled
Control algorithm separately by the regenerative and friction brakes, is developed. Simulations of the proposed control
Hardware-in-the-loop validation strategy are conducted based on system models built using MATLAB/Simulink. The simulation results
demonstrate that the developed strategy enables improved control in terms of vehicle stability and
braking performance under different emergency driving conditions. To further verify the synthesized
control algorithm, hardware-in-the-loop tests are also performed. The experimental results validate
the simulation data and verify the feasibility and effectiveness of the developed control algorithm.
Ó 2015 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.enconman.2015.09.062
0196-8904/Ó 2015 Elsevier Ltd. All rights reserved.
C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529 521
Nomenclature
of electric motors and batteries. Hence, the friction brake is also 2. System dynamics modeling
required to supplement the overall brake demand. In [14], an
electrically-controlled regenerative braking system based on the 2.1. System layout of the case-study vehicle
conventional hydraulic ABS techniques was developed, and a con-
trol algorithm coordinating the braking energy regeneration and Fig. 1 illustrates the diagram of the proposed electric vehicle
ABS control functions was designed. In [15], three executive con- equipped with regenerative braking system and pneumatic brake
trol strategies, namely, the ‘‘PQ-method,” ‘‘frequency selection by system. A brake control unit takes over the control of all the four
filter,” and ‘‘model following control” strategies were proposed. modulating valves, and communicates with the motor control unit
As the electric motor responds more accurately and rapidly than and vehicle control unit via controller area network (CAN) bus.
the friction brake, optimal ABS control is expected. Simulations During normal decelerations, the electric motor works as a gener-
and road tests validated the control performance of the proposed ator, generating electrical energy and charging the battery [18]. In
method. the meantime, the electric motor provides regenerative braking
Nevertheless, all the existing research is mainly based on pas- torque, which is transmitted by the drivetrain, and finally exerted
senger cars equipped with hydraulic ABS, and the coordinated con- on wheel, decelerating the vehicle.
trol of regenerative brake and pneumatic ABS under vehicle’s However, when the vehicle enters a critical driving situation, if
emergency braking conditions has rarely been studied. Since the the wheel has a locked tendency which is detected by the wheel
dynamic behavior of a pneumatic braking system is quite distin-
guished from a hydraulic one [16,17], the cooperation of regenera-
tive braking system and pneumatic ABS is worthwhile researching
for improving advanced control technologies of electric vehicles.
In this paper, a novel control algorithm coordinating the regen-
erative braking system and pneumatic braking system for an elec-
trified bus under ABS control procedures has been proposed. In
Section 2, the system layout of an electrified bus equipped with a
regenerative brake system has been introduced. And the models
of vehicle dynamics and main components related to the regener-
ative and the frictional brakes of the case-study electric bus have
been built in MATLAB/Simulink. In Section 3, the target brake tor-
que of antilock braking control was obtained based on the phase
plane theory. And then, according to the different system charac-
teristics of regenerative and friction brakes, a novel ABS control
algorithm was designed wherein the brake torque command is
divided into steady and dynamical parts that are handled sepa-
rately by the pneumatic and regenerative brakes. In Sections 4
and 5, the control strategy was evaluated using simulations and
hardware-in-the-loop (HIL) tests and compared with two baseline
control strategies. Finally, some concluding remarks are provided. Fig. 1. System layout of the electric vehicle equipped with regenerative braking.
522 C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529
sM T_ M þ T M ¼ T M;ref ð1Þ
speed sensor, then the presented safety control program will be
activated. The brake control unit will regulate the pneumatic where TM is the actual value of motor torque, and TM,ref is the motor
pressure by controlling the modulating valves and gives out a reference torque.
motor torque command to motor control unit in the meantime. The output torque of electric motor will then transmitted by the
The motor control unit controls the electric motor working as a gearbox, distributed by the differential and final drive unit, and
generator, exerting regenerative braking torque via transmission finally exerted on the two driven wheel. Thus, during a regenera-
on the drive-axle as required. Thus, brake blending is carried out tive deceleration process, the regenerative braking torque on a dri-
by combining the regenerative and pneumatic braking torques. ven wheel can be calculated by [20]:
These hybrid brakes work cooperatively with each other to avoid
i0 ig
wheel locking and to maintain the stability of the vehicle, and Tb regen ¼ TM ð2Þ
the regeneration efficiency is not the first priority under those cir-
2
cumstances. Some key parameters of the case-study electrified where ig is the transmission ratio of the gearbox, and i0 is the final
vehicle are listed in Table 1. drive ratio.
In the case-study electric vehicle, the regenerative braking The battery model is based on the model of the nickel–metal
torque of the braking energy regeneration system is provided by hydride (NiMH) batteries of a General Motors electric vehicle in
where m _ c is the time rate of change of the air mass in the control
volume, and q is the density of air inside the control volume. uc is 2.5. Vehicle model
the velocity of air at the exit section of the nozzle, which is the
velocity of air entering the control volume, and Ap is the cross- The model of vehicle dynamics has been built in MATLAB-
sectional area of the valve opening. Simulink, and has been validated by experiments [12]. The OXYZ
The velocity of air can be expressed as [21]: coordinate system and the vehicle model, shown in Fig. 4, have
" ððc1Þ=cÞ #!1=2 its origin at the center of gravity of the vehicle.
2c Po Pb The longitudinal motion of the vehicle is expressed as
uc ¼ C d 1 sgnðPo Pb Þ ð5Þ
c 1 qo Po
mðu_ v cÞ ¼ ðF x11 þ F x12 Þ cos d ðF y11 þ F y12 Þ sin d þ F x21
where Cd is a discharge coefficient compensating for the losses dur- CDA
þ F x22 ð3:6uÞ2 ð8Þ
ing the flow and taken as 0.7. 21:15
The signum function above is defined as: where m is the overall mass of the vehicle, CD is the coefficient of
Po Pb the air resistance, A is the frontal area. d is the steering angle of
sgnðPo Pb Þ ¼ ð6Þ the front wheels, u is the velocity along OX, and v is the velocity
jP o Pb j
along OY, Fx is the longitudinal force of the wheel, Fy is the trans-
Regard the air in the control volume as ideal gas, a function as verse force of the wheel, Fz is the vertical force of the wheel.
follows can be obtained by: The motion of the tyre of a drive wheel can be expressed as
_
1 Pb V c Pb V_ c 1
_c¼
m þ ð7Þ J w þ J pt x_ ij ¼ T dij T bij þ F xij r; ði ¼ 2; j ¼ 1; 2Þ ð9Þ
c RT c RT c 2
where Vb is the volume of the control volume, Tc is the temperature where Jw is the moment of inertia of tyre, Jpt is the moment of iner-
of the control volume, R is the gas constant for air, and c is the ratio tia of the powertrain which is shared by the two drive wheels, xij is
of the specific heats. the rotational speed of the wheel, r is the nominal radius of the tyre,
Tdij and Tbij are the driving and braking torques exerted on wheel
respectively, ij = 11 for the front left wheel, ij = 12 for the front right
wheel, ij = 21 for the rear left wheel, and ij = 22 for the rear right
wheel.
The slip ratio S of the braking wheel is the key parameter for
vehicle control under safety–critical driving situations, and conver-
gence to an ideal slip ratio S⁄ would lead to the maximum brake effi-
ciency and brake stability of the vehicle. To achieve the ideal slip
ratio, a first-order system is built based on the phase plane theory
_ as shown in Eq. (10):
using the wheel slip ratio S and its derivative S,
S_ ¼ kðS S Þ ð10Þ
Fig. 3. Diagram of the control volume of the pneumatic brake system.
524 C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529
(1) S ultimately converges to S⁄. 3.3. Control allocation between regenerative and friction brakes
When S > S⁄, the locus is on the upper half-plane. The locus In an electric bus equipped with a regenerative braking system,
moves toward the right and S increases until S = S⁄. Conversely, the deceleration command can be executed through the coopera-
on the opposite side, when S < S⁄, the locus moves toward the left tive operation of the regenerative and friction brakes. Therefore,
until S = S⁄. the distribution of the target brake torque between the motor
and friction torques should be studied.
(2) The system responses rapidly and with minimal oscillation. The characteristics of the regenerative and pneumatic brakes
are far distinguished according to their system dynamics.
_ is large
_ is positively correlated with |S S⁄|. jSj The motor torque can satisfy the high frequency demand of the
In the system, jSj
target brake torque of the vehicle owing to its rapid response and
when S is far from S⁄, thereby quickly driving S toward S⁄. When S
accurate control. Additionally, the electric motor can operate in
_ become small, and hence, S gradually converges
approaches S⁄, jSj either the brake or drive mode, which provides much more flexibil-
to S⁄. ity than the pneumatic brake in dynamics control.
In conclusion, for the system described by Eq. (1), the wheel slip While in the pneumatic brake system, solenoid valves are used
ratio converges to the ideal value at a high speed and with minimal to regulate the braking pressure during the ABS control procedure.
oscillation. As a result, the available tyre–road adhesion is fully Brake pressure can be increased, decreased, or held through the
used to achieve the maximum brake force possible. pulse width modulation of the valves. The brake force cannot be
modulated quickly and accurately enough owing to the existing
3.2. Optimal braking torque time delay between the valve action and the pressure variation
in the brake line. Therefore, the pneumatic brake is unsuitable
To achieve the expected relationship between S and S⁄ pre- for supplying the high frequency demand of the brake torque.
sented in Eq. (10), the optimal brake torque of the wheel needs Based on the different dynamics of the pneumatic and regener-
to be figured out. Considering the drive wheel as the control object ative brakes analyzed above, the target optimal brake torque T b is
on which regenerative and pneumatic brake forces are applied, the divided into two parts, as shown in Eq. (16). They are Tsteady and
dynamic model restricted to longitudinal motion is illustrated in Tdyna, which are the steady and dynamical parts, respectively, of
Fig. 6. the braking torque command.
The governing equations for the motions of the drive wheel can
T b ¼ T steady þ T dyna ð16Þ
be given by:
The steady part of the brake torque demand is calculated based
Fx
u_ ¼ ð11Þ on the current wheel–road adhesion that keeps the slip ratio
mt stable. Defining that, when the brake torque equals Tsteady, S will
remain unchanged, as shown in Eq. (17).
1
x_ ¼ ðr F x þ T b Þ ð12Þ _
_ SÞ r x
uð1
I S_ ¼ ¼0 ð17Þ
u
where mt is the wheel’s vertical load, r is the wheel radius, and I is
Combining Eqs. (11), (12), (16) and (17), the steady part of the brake
the wheel’s inertia.
demand Tsteady, which the friction braking system is required to be
in charge of, can be expressed as follows:
Iu_
Tb fric ¼ T steady ¼ ð1 SÞ þ mt r u_ ð18Þ
r
When S departs from S⁄, Tdyna is required to provide the dynam-
ical part of the brake torque demand. It combines with Tsteady,
enabling the total torque to achieve the target value and the slip
ratio to converge to the ideal value.
Therefore, by combining Eqs. (15)–(18), the target regenerative
braking torque, which is provided by the electric motor and should
take over the dynamical part of the overall brake demand, can be
Fig. 6. Longitudinal dynamical model of the wheel during deceleration. expressed as follows:
C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529 525
Fig. 7. Simulation results of a brake process on an ice road. Fig. 8. Simulation results of a brake process on a jump-l road.
526 C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529
5. Hardware-in-the-loop tests
Table 2
Comparison of three control strategies in simulation.
Road condition Strategy Mean deceleration (m/s2) Stopping distance (m) Performance improvement (%)
Ice Conventional ABS 0.71 170.8 16.90
Integrated RBS/ABS 0.74 165.7 12.16
Proposed control 0.83 150.9 –
Snow Conventional ABS 1.40 86.1 16.43
Integrated RBS/ABS 1.47 84.5 10.88
Proposed control 1.63 78.0 –
Gravel Conventional ABS 5.03 28.9 3.38
Integrated RBS/ABS 5.10 27.9 1.96
Proposed control 5.20 26.7 –
Jump-l Conventional ABS 1.60 66.9 15.63
Integrated RBS/ABS 1.81 59.5 2.21
Proposed control 1.85 58.7 –
Split-l Conventional ABS 2.35 57.8 8.94
Integrated RBS/ABS 2.37 56.2 8.02
Proposed control 2.56 53.6 –
C. Lv et al. / Energy Conversion and Management 106 (2015) 520–529 527
are sent from vehicle control unit, indicating that the ABS function HIL-tested separately on the ice, snow, gravel, jump-l road and
or regenerative braking function is activated, respectively. Fig. 12 is split-l road, using same scenarios and procedures by the present
a photograph of the developed HIL bench test system. authors [21]. And the HIL test results of the newly developed con-
trol on the ice, jump-l road, and split-l are shown as follows in
Fig. 13, Fig. 14 and Fig. 15, respectively.
5.2. Test results and analysis
Under the proposed control, when the wheel tends to lock, the
brake controller regulates the pressure in the wheel cylinder with a
Two existing approaches, i.e. the pure pneumatic ABS control
low frequency. Additionally, the motor switches in between the
and the integrated regenerative braking control have been
regenerative and the drive modes and quickly provides a regener-
ative braking torque or a driving torque as required. The overall
requirements of ABS control are satisfied. No wheel lock occurs
during the entire deceleration process. From the above experimen-
tal results, it is observed that the angular speed of the locked wheel
recovers rapidly enough. Thus, the deceleration of the vehicle is
maintained appropriately.
Table 3
HIL test results of the proposed control with two conventional strategies.
6. Conclusion [5] Zhang J, Lv C, Gou J, et al. Cooperative control of regenerative braking and
hydraulic braking of an electrified passenger car. Proc Inst Mech Eng, Part D: J
Automob Eng 2012;226(10):1289–302.
In this paper, a novel control strategy of braking energy regen- [6] Gao Y. Electronic braking system of EV and HEV-integration of regenerative
eration system for electric vehicles during safety–critical driving braking, automatic braking force control and ABS. SAE Technical Paper 2001-
01-2478; 2001.
maneuvers is proposed. The models of vehicle dynamics and main
[7] Katrašnik T. Analytical framework for analyzing the energy conversion
components related to the regenerative and the frictional hybrid efficiency of different hybrid electric vehicle topologies. Energy Convers
brakes of the electric bus were built in MATLAB/Simulink. Based Manage 2009;50(8):1924–38.
[8] Tie SF, Tan CW. A review of energy sources and energy management system in
on the phase plane theory, the target brake torque of antilock con-
electric vehicles. Renew Sustain Energy Rev 2013;20:82–102.
trol was obtained to ensure full utilization of road adhesion. [9] Maxoulis CN, Tsinoglou DN, Koltsakis GC. Modeling of automotive fuel cell
According to the different system dynamics of regenerative operation in driving cycles. Energy Convers Manage 2004;45(4):559–73.
and friction brakes, a novel hybrid ABS control algorithm was [10] Lv C, Zhang J, Li Y, et al. Synthesis of a hybrid-observer-based active controller
for compensating powetrain backlash nonlinearity of an electric vehicle during
designed wherein the brake torque command is divided into regenerative braking. SAE Int J Altern Powertrains 2015;4(1):190–8.
steady and dynamical parts that are handled separately by the [11] Von Albrichsfeld C, Karner J, Brake system for hybrid and electric vehicles, SAE
pneumatic and regenerative brakes. The control strategy was eval- Technical Paper, 2009-01-1217; 2009.
[12] Zhang J, Lv C, Qiu M, et al. Braking energy regeneration control of a fuel cell
uated using simulations and HIL tests and compared with two hybrid electric bus. Energy Convers Manage 2013;76:1117–24.
baseline control strategies. [13] Tur O, Ustun O, Tuncay R. An introduction to regenerative braking of electric
The results of the simulations and HIL tests indicated that the vehicles as anti-lock braking system, Intelligent Vehicles Symposium, IEEE;
2007. p. 944–8.
proposed control strategy yields better braking performance and [14] Zhang J, Lv C, Yue X, Qiu M, et al., Development of the electrically-controlled
stability than the conventional ABS control and the integrated regenerative braking system for electrified passenger vehicle, SAE Technical
regenerative braking control with ABS owing to the reasonable Paper 2013-01-1463; 2013.
[15] Hori Y. Future vehicle driven by electricity and control-research on
coordination between the friction and regenerative brakes. Based
four-wheel-motored” UOT Electric March II”. Ind. Electron., IEEE Trans.
the HIL test results, mean deceleration of the vehicle on ice road 2004;51:954–62.
has been improved by 13.89% and 9.30% respectively compared [16] Junzhi Z, Yutong L, Chen L, et al. New regenerative braking control strategy for
rear-driven electrified minivans. Energy Convers Manage 2014;82:135–45.
to the two benchmarks. The simulation and HIL experiment results
[17] Huang K, Quang K, Tseng K. Study of recycling exhaust gas energy of hybrid
for different braking conditions also demonstrate the adaptability pneumatic power system with CFD. Energy Convers Manage 2009;50
and feasibility of the proposed control strategy. (5):1271–8.
[18] Çağatay Bayindir K, Gözüküçük MA, Teke A. A comprehensive overview of
hybrid electric vehicle: powertrain configurations, powertrain control
techniques and electronic control units. Energy Convers Manage 2011;52
Acknowledgement (2):1305–13.
[19] Lv C, Zhang J, Li Y. Extended-Kalman-filter-based regenerative and friction
The authors would like to thank the Natural Science Foundation blended braking control for electric vehicle equipped with axle motor
considering damping and elastic properties of electric powertrain. Vehicle
of China [project no. 51475253] for funding this work. Syst Dyn 2014;52(11):1372–88.
[20] Lv C, Zhang J, Li Y, et al. Hardware-in-the-loop simulation of pressure-
difference-limiting modulation of the hydraulic brake for regenerative braking
References control of electric vehicles. Proc Inst Mech Eng, Part D: J Automob Eng
2014;228(6):649–62.
[21] Zhang J, Kong D, Lv C, et al. Optimization of control strategy for regenerative
[1] Katrašnik T. Energy conversion phenomena in plug-in hybrid-electric vehicles.
braking of an electrified bus equipped with an anti-lock braking system. Proc
Energy Convers Manage 2011;52(7):2637–50.
Inst Mech Eng, Part D: J Automob Eng 2012;226(4):494–506.
[2] Nguyen AT, Lauber J, Dambrine M. Optimal control based algorithms for
[22] Khalil H. Nonlinear systems. 3rd ed. Prentice Hall; 2001.
energy management of automotive power systems with battery/
[23] ABS/ASR ‘‘D”-‘‘Cab” – Version anti-lock braking system for commercial
supercapacitor storage devices. Energy Convers Manage 2014;87:410–20.
vehicles, 1st ed. Brussels: WABCO; 2006.
[3] Park SH, Lee CS. Applicability of dimethyl ether (DME) in a compression
[24] Palladino A, Fiengo G, Lanzo D. A portable hardware-in-the-loop (HIL) device
ignition engine as an alternative fuel. Energy Convers Manage
for automotive diagnostic control systems. ISA Trans 2012;51:229–36.
2014;86:848–63.
[25] Pugi L, Galardi E, Carcasci C, Rindi A, Lucchesi N. Preliminary design and
[4] Lv C, Zhang J, Li Y, et al. Mechanism analysis and evaluation methodology of
validation of a Real Time model for hardware in the loop testing of bypass
regenerative braking contribution to energy efficiency improvement of
valve actuation system. Energy Convers Manage 2015;92:366–84.
electrified vehicles. Energy Convers Manage 2015;92:469–82.