Proceedings of 2015 IEEE

International Conference on Mechatronics and Automation

August 2 - 5, Beijing, China

Modeling and Analysis of Regenerative Braking System

for Electric Vehicle Based on AMESim
Junzhi Zhang, Ye Yuan, Chen Lv and Yutong Li
State Key Laboratory of Automotive Safety and Energy
Tsinghua University
Beijing 100084, China
[email protected]

Abstract – This paper presents a parametric and since these systems are more advantageous over the parallel
modularized method in modeling the hydraulic components of ones considering the regeneration efficiency and the braking
the regenerative braking system, through which the newly built feel [4]. Toyota, TRW and Honda have developed serial type
model has gained noticeable improvement in precision. The systems based on electronic hydraulic brake (EHB) systems,
strategy for the exiting of the regenerative braking is optimized.
which have been implemented successfully on
The simulations are conducted in the co-simulation platform
between AMESim and MATLAB/Simulink at the initial vehicle commercialized HEVs including Toyota Prius, Ford Escape,
speed of 40 Km/h. The simulation results show that the newly and Honda Insight [5][6][7]. The electric vehicle Leaf adopted
built hydraulic model can describe the process of pressure a newly designed brake system named the Electrically-Driven
increase and decrease precisely. Meanwhile the motor cooperates intelligent Brake (EDiB) developed by Hitachi. In the EDiB
with the hydraulic braking system well throughout the whole system, a redesigned master cylinder was involved, and the
braking procedure. The maximum jerk exerted on the vehicle is brake pressure is adjusted by an electrically driven motor
decreased from 2.69 m/s3 to 0.59 m/s3 during the exiting of through a ball screw [8].
regenerative braking, and the regeneration efficiency is increased The present authors have been dedicated to the research
to 76.18%.
and development of regenerative braking for a long period and
Index Terms – regenerative braking, modeling of hydraulic
made some progress [9][10][11][12]. In order to lay
braking system, co-simulation, brake comfort, regeneration foundation for developing of the next generation of
efficiency regenerative braking system, to provide basis for parameter
tuning and to shorten the time cost from proto design to
I. INTRODUCTION application, more precise hydraulic system models and
flexible simulation methodology are needed.
Regenerative braking, as a noticeable advantage of
In this study, AMESim is adopted due to its superiority in
electrified vehicles, can significantly improve the energy
parametric and modularized modeling of hydraulic
consumption and satisfy the social need of environment
components such as pumps, pump motors, solenoid valves,
conservation. This technology offers the capability to recover
accumulators etc. Most of the proto models have passed the
kinetic energy during decelerations. Therefore, plenty of
industrial verification [13]. Along with the parametric and
research and development has been conducted worldwide
modularized modeling capability, AMESim cooperates with
among automakers, parts manufacturers and researchers,
MATLAB/Simulink by working in co-simulation mode. Co-
which were mainly concentrated on the braking force
simulation renders the integrator selection and simulation step
distribution strategies and the scheme design of regenerative
setting more flexible.
braking systems [1][2].
In terms of braking force distribution strategies, Luo and II. SYSTEM CONFIGURATION
Li et al. designed a distribution model of the regenerative
A. System layout
brake and the frictional brake based on optimal control theory.
A newly developed regenerative brake system named
The simulation result shows that the proposed control strategy
EABS is proposed in [10]. The main component of EABS is
can speed up the dynamic response to braking and the energy
the EABS hydraulic control unit, which is mounted between
regeneration efficiency can be increased by around 10% [3].
the master cylinder and the wheel cylinders in the brake line,
Zhang and Lv et al. at Tsinghua University have proposed
taking the place of brake control unit of conventional anti-lock
three different braking force distribution strategies, maximum-
braking system (ABS) as shown in Fig. 1. The EABS is
regeneration-efficiency strategy, good-pedal-feel strategy and
developed on basis of conventional ABS by adding 6 high
coordination strategy. The results of simulations and road tests
speed solenoid valves and 2 pedal stroke simulators. The
show that the maximum-regeneration-efficiency strategy could
brake pressure is still built up under the assistance of a
hardly be adopted due to terrible brake pedal feeling. The
vacuum booster.
good-pedal-feel and coordination strategies are advantageous
with respect to brake comfort and regeneration efficiency,
ensuring the stability and brake safety of the vehicle [4].
In terms of system scheme design, the serial type
regenerative braking systems have been widely developed,

978-1-4799-7098-8/15/$31.00 ©2015 IEEE 1307

 The theoretical foundation of pressure increase and
decrease are set up as shown in Fig. 3, to verify the models
built in AMESim. In Fig. 3, pm denotes the entrance pressure
of the inlet valve, p0 denotes the exit pressure of the outlet
valve, i.e. the pressure in low-pressure accumulator, pw
denotes the pressure of the wheel cylinder, k is the stiffness
of the spring in wheel cylinder, and x denotes the spring
displacement.

Fig. 1 Overall structure of the regenerative braking system

EABS hydraulic control unit is composed of two parts,

i.e. the braking control unit (BCU) and the hydraulic pressure
modulator as shown in Fig. 2. The functions of ABS and
regenerative braking are integrated into the BCU, which
communicates with the Vehicle Control Unit (VCU) via CAN
bus and drives the hydraulic pressure modulator to adjust the
brake pressure rapidly and precisely as demand.

Fig. 3. Diagram of the hydraulic braking system model.

The hydraulic fluid flow of a valve can be expressed as

follow:
2 ⋅ Δp
Q = Cd A , (1)
ρ
where Cd denotes the flow coefficient of the valve and is set
as 0.7; A denotes the equivalent cross sectional area at
maximum opening and is set as 0.33 mm2, ρ denotes the
density of the hydraulic fluid, and Δp denotes the pressure
difference between the inlet and outlet of the valve.
For pressure increase process,
Fig. 2 EABS hydraulic control unit Δp = p m − p w . (2)
For pressure decrease process,
B. System modeling Δp = pw − p0 . (3)
A control system of regenerative braking needs to be For wheel cylinder,
established for the simulation and analysis of the regenerative Qdt = π rw2 dx , (4)
braking procedure. In this study, the model of vehicle where rw2 denotes the diameter of the wheel cylinder piston.
dynamics, the tires, the electric motor, the battery and the
braking control unit are set up in MATLAB/Simulink. The According to the equilibrium criteria,
model of the whole hydraulic brake system is set up in π rw2 dpw = kdx . (5)
AMESim simultaneously. The detailed methods for modeling By combing (4) and (5), the rate of pressure change in
the vehicle dynamics and the tire have been described in [9]. wheel cylinder can be expressed as follows:
The modeling of the motor and the battery is based on the dpw k 2 ⋅ Δp
foundation of [12]. The hydraulic brake system is modeled in = 2 4 Cd A . (6)
AMESim and the procedure is expounded as follows.
dt π rw ρ
The modeling procedure of the hydraulic braking system Valve model is built based on the physical structure of
in AMESim is composed of two parts, i.e. valve modeling and high-speed switch valves in AMESim as shown in Fig. 4. The
pipeline construction. parameters of the valve is set based on an on-off solenoid

1308
valve supplied by Zhejiang Asia-Pacific Mechanical & built in AMESim. The simulation result is presented in Fig. 6.
Electronic Co., Ltd as shown in Table I. In first 2 seconds, only the inlet valve (normally open) is open
while in last 2 seconds only the outlet valve (normally closed)
Pushrod is activated. The entrance pressure of the inlet valve maintains
Electromagnetic coil Valve core at 3 MPa during this whole simulation procedure.
Return spring The simulation result indicates that the model in AMESim
has taken the opening time and closing time of the valve into
consideration, so the curve climbs slowly at the beginning
than the other and needs more time to drain the wheel
Inlet cylinder.
PWM control signal
Outlet
TABLE I
Fig. 4 The valve model in AMESim
THE PARAMETERS OF THE VALVE

The pipeline construction is based on Fig. 1, and Parameter Valve

AMESim has provided a more physical modeling method than seat diameter (hole) 0.65 mm
MATLAB/Simulink. The pumps, pump motor, clippers, seat semi-angle 57°
reservoirs and accumulators are chosen from AMESim
prototypical library, and are set with proper parameters. The ball diameter 1.588 mm
overall hydraulic brake system is shown in Fig. 5. The maximum displacement of core 0.3 mm
MATLAB/Simulink interfaces are also modeled to fulfill the
spring stiffness 0.34 N/mm
co-simulation between MATLAB/Simulink and AMESim.
C. Model validation
Simulations are conducted in MATLAB/Simulink and
AMESim separately to test the validity of the valve model

Fig. 5 The overall hydraulic brake system in AMESim

1309
 B. Modified regeneration brake exiting strategy
For the baseline regenerative brake exiting strategy, at the
exiting stage of the regenerative brake, the motor torque
evacuates at the maximum allowed speed (500 Nm/s),
meanwhile the inlet valves of the front wheel cylinders keep
the maximum opening state to supply the deficiency of total
brake force. This strategy will cause the sinking of the brake
pedal and jerks.
For the modified exit strategy, evacuating speed of the
motor torque is calculated based on the difference of master
cylinder and the front wheel cylinders. Another compile of
PID controller parameters and lookup table is set up as shown
in Fig. 8.

Main cylinder
Fig. 6 Valve models simulation result
+-
III. REGENERATION BRAKE CONTROL STRATEGY
Regenerative -
+
Regeneration control strategy is composed of hydraulic brake exiting
control strategy e
pressure modulation algorithm and braking force distribution PID
Controller
strategy. Hydraulic pressure modulation algorithm is based on
PID controller and lookup tables which have been introduced k1
++
Lookup
minutely in [12], but jerks happen at the end of regeneration Table
k2 Valve_ CMD
brake process and the potential of the motor is not fully
Treg_cmd EABS Hydraulic
exploited. In this study, a modified regenerative braking exit Pressure Modulator
strategy is proposed and the baseline strategy is also Electric motor
Pw_act
introduced as a contrast. Treg_act
Regenerative
A. Braking force distribution strategy braking Hydraulic braking
(exiting)
As shown in Fig. 7, a constant brake force imitating the
engine brake of conventional vehicles, and the front brake
force is divided into 2 parts. At the beginning of the brake, the Total brake
brake force demand of front axle can be satisfied by motor
Fig. 8 Modified regenerative brake exiting algorithm
generation torque, i.e. regenerative braking. When the brake
demand of front axle hits the motor limit, the hydraulic brake
force is exerted. The whole brake force of rear axle is IV. VALIDATION OF CONTROL STRATEGY
achieved only by hydraulic brake force.
The brake force distribution (BFD) is identical with the Co-simulations are conducted via the co-simulation
BFD of conventional vehicles, which can be seen in Fig. 7. interface between MATLAB/Simulink and AMESim for
The brake force of rear axle is proportional to the sum of validating the effectiveness of the regenerative brake control
regenerative brake force and the hydraulic brake force of front strategy. The results are focused on the regeneration efficiency
axle. and the cooperative effects of hydraulic brake and the
regenerative brake.
The co-simulation is conducted under normal urban
braking condition, where the initial vehicle speed is 40 km/h,
the adherence coefficient is 0.8, and the master cylinder
pressure is set as a ramp increase and stabilizes at 3 MPa.
The baseline control strategy simulation results are shown
in Fig. 9. The maximum regenerative braking torque is 120.41
Nm, which is slightly lower than the electric motor torque
limit (145Nm), thus indicating high regeneration efficiency.
The total energy regenerated is 57.8 KJ, which is 75.84% of
the total recoverable energy. The duration for the motor
regeneration torque exiting is 0.24 s, during which the
maximum jerk is 2.69 m/s3, demonstrating good braking
Fig. 7 Distribution between regenerative brake and hydraulic brake comfort. The maximum equivalent front-left wheel cylinder

1310
 76.21 KJ

57.8 KJ

(a) Diagram of distribution of different pressures (b) Diagram of energy transformation

12.5 m/s3
0.24 s REG exiting
2.69 m/s3

120.41 Nm
-11.05 m/s3

(c) Diagram of jerk, deceleration and vehicle speed (d) Diagram of motor torque and battery charge power
Fig. 9 Simulation results of the baseline strategy
pressure is 3.05 MPa, indicating the validity of the and the curve representing the jerk exerted on the vehicle is
regeneration brake control strategy. significantly smoother than that of baseline regenerative brake
The simulation results for the modified regenerative brake control strategy. The energy recovered is 76.18% of the total
exit strategy are illustrated in Fig. 10. The main difference is recoverable energy, demonstrating that a slight improvement
the magnitude of the jerk during motor torque exit procedure. in regeneration efficiency has achieved due to the duration
The time for the motor regeneration torque exiting is 0.48 s extension of the motor regenerative brake exiting.

3.05MPa

3.00MPa 76.10 KJ

57.97 KJ

0.05MPa

(a) Diagram of distribution of different pressures (b) Diagram of energy transformation

0.48s REG exiting

7.25 m/s3

0.59 m/s3
120.44Nm

3
-8.49 m/s

(c) Diagram of jerk, deceleration and vehicle speed (d) Diagram of motor torque and battery charge power
Fig. 10 Simulation results of the modified strategy

V. CONCLUSION

1311
 In order to enhance the precision of simulation models brake for regenerative braking control of electric vehicles,” Proceedings
of the Institution of Mechanical Engineers, Part D: Journal of
and shorten the time cost from proto design to application, a Automobile Engineering 228.6 (2014): 649-662.
new parametric and modularized modeling method was [12] J. Zhang, Y. Li, C. Lv, and Y. Yuan, “New regenerative braking control
researched with AMESim and MATLAB/Simulink software strategy for rear-driven electrified minivans,” Energy Conversion and
packages. The exiting strategy was optimized to reduce jerks Management 82 (2014): 135-145.
[13] A. Lynn, E. Smid, M. Eshraghi, N. Caldwell, and D.Woody, “Modeling
and take full advantage of the motor braking torque. hydraulic regenerative hybrid vehicles using AMESim and
Simulations were carried out on models built via co- Matlab/Simulink,” Proceedings of SPIE, the International Society for
simulation interfaces between MATLAB/Simulink and Optical Engineering. Society of Photo-Optical Instrumentation
AMESim. The results of simulations show that the newly built Engineers, 2005.
hydraulic model can describe the process of pressure increase
or decrease precisely, the motor cooperates with the hydraulic
braking system well throughout the whole braking procedure.
The total time consumption for the exiting of regenerative
braking is prolonged from 0.24 s to 0.48 s. Hence the
maximum jerk exerted on the vehicle during the exiting of
regenerative braking is decreased from 2.69 m/s3 to 0.59 m/s3,
and the regeneration efficiency is increase slightly to 76.18%.
Further studies will be carried out in some areas such as
the following: modeling of the powertrain considering
dynamic behaviors, modeling of the master cylinder
considering pedal feel, optimization and matching parameters
of hydraulic components such as the reservoir and pump etc.
ACKNOWLEDGMENT
This paper is supported by the Natural Science
Foundation of China [project No. 51475253] and National
Key Technology Research and Development Program of the
Ministry of Science and Technology of China [project No.
2013BAG08B01].

REFERENCES
[1] Y. Gao and E. Mehrdad, “Electronic braking system of EV and HEV-
integration of regenerative braking,” Automatic Braking Force Control
and ABS. SAE paper 2001-01-2478, 2001.
[2] O. Tur, O. Ustun, and R.N. Tuncay, “An introduction to regenerative
braking of electric vehicles as anti-lock braking system,” Intelligent
Vehicles Symposium, IEEE, 2007.
[3] J. Zhang, C. Lv, J. Gou, and D. Kong, “Cooperative control of
regenerative braking and hydraulic braking of an electrified passenger
car,” Proceedings of the Institution of Mechanical Engineers, Part D:
Journal of Automobile Engineering 226.10 (2012): 1289-1302..
[4] Y. Luo, P. Li, D. Jin, and K. Li, “A study on regenerative braking
strategy based on optimal control theory,” Automot. Eng 28.4 (2006):
356-360.
[5] Y. Aoki, K. Suzuki and H. Nakano, “Development of hydrodynamic
servo brake system for cooperative control with regenerative brake,”
SAE paper 2007-01-0868, 2007.
[6] E. Nakamura, M. Soga, A. Sakai, A. Otomo, et al., “Development of
Electronically Controlled Brake System for Hybrid Vehicle,” SAE
Technical Paper 2002-01-0300, 2002.
[7] G. Blaise, and A. Ann, “Slip Control Boost Control System,” US Patent
Application, 20090077963A1, 2009.
[8] Y. Ohtani, T. Innami, T. Obata, T. Yamaguchi, et al., “Development of
an Electrically-Driven Intelligent Brake Unit,” SAE Technical Paper
2011-01-0572, 2011.
[9] J. Zhang, X. Chen, P. Zhang. “Integrated control of braking energy
regeneration and pneumatic anti-lock braking,” Proc IMechE Part D: J
Automob Eng 2010;224:587–610.
[10] J. Zhang, C. Lv, X. Yue, M. Qiu, J. Gou, and C. He, “Development of
the electrically-controlled regenerative braking system for electrified
passenger vehicle,” SAE Technical Paper, 2013-01-1463. 2013.
[11] C. Lv, J. Zhang, Y. Li, D. Sun, and Y. Yuan, “Hardware-in-the-loop
simulation of pressure-difference-limiting modulation of the hydraulic

1312