Regenerative braking-driving control system - IEEE Xplore

Regenerative Braking-Driving Control System 1

1

1

Yu-Chan Chen 1 , Yu-Chen Chang , Jiang-Feng Cheng 1,2 , Wen-Cheng Yu and Chun-Liang Lin 1. Department of Electrical Engineering, National Chung Hsing University, Taichung, Taiwan 2. School of Automation Science and Electrical Engineering, Beihang University, Beijing, China Abstract—Currently, electric vehicle technology is becoming more and more mature. Although the anti-lock braking system (ABS) has been commonly applied, most electric vehicles (EVs) still use traditional hydraulic-based disc braking, which has the drawbacks that vehicle wheels are is easy to skid in the rainy day, and easy to be abraded during emergency brake. As a novel method of braking, regenerative braking has the advantages of compact structure, sensitive response, reliability and controllable braking distance. In this research task, a regenerative driving and braking control system for EVs with satisfactory braking performance is proposed. When braking, a motor is converted into a generator-the acquired energy can be used to generate reverse magnetic braking torque with fast response. On this basis, an anti-lock braking controller is realize. A PID controller is also designed to drive the motor and a fuzzy slip ratio controller is designed and used to obtain the optimal slip ratio. Finally, real-world experiments are conducted to verify the proposed method. Index Terms—Integrated driving and braking scheme, ABS, regenerative braking, fuzzy control.

T

I.

INTRODUCTION

he electric vehicle braking issue, especially the electromagnetic brake technology, has attracted much attention all over the world. Several braking methods have been proposed, including brushless motor controller, regenerative braking, under-bridge short circuit braking, and integrated driving and braking control system. The works [1-5] investigate the development of brushless motor controller and use dynamic braking to stop the motor. The regenerative braking system was developed to charge the battery to extend the vehicle traveling range [6-11]. The volt-second balance and Ampere-second balance methods were proposed to investigate the relationship between duty cycle and braking torque [12]. With regard to the short circuit braking design, the under-bridge short circuit scheme makes all the back electromotive force (back-EMF) to be consumed in the motor coil. As the back-EMF is fully consumed, the braking torque is greater than those of the usual dynamic braking and regenerative braking. In addition, ABS design based on the short circuit braking was incorporated in a series of research tasks in [13-16]. In [17], an integrated driving and braking control system based on the short circuit brakes is further proposed through combining the braking system with the driving system. In view of the investigation above, regenerative braking can no longer recharge the battery if the battery has already been fully charged, which may weaken the braking torque. Thus, to solve the drawback, the integrated driving and braking system is proposed, which changes the load of regenerative braking from the battery to a short circuit. However, it is found that this

c 978-1-5386-3758-6/18/$31.00 2018 IEEE

method cannot stop the vehicle immediately on road (i.e., the braking torque is not large enough). Based on the above-mentioned research works, this paper first proposes a method by developing a reversal magnetic field with back-EMF to enhance the braking effect. With the anti-lock braking system (ABS) operational rules, the braking torque of reverse braking design is controlled to drive the slip ratio falling within the range to ensure the best road adhesion during emergency stop. Extensive real-world experiments have been conducted to verify the proposed approach. II.

THE ELECTRIC MACHINES

Six-step commutation and Hall sensor are used to control the motor in this research task. Electronic switches realized by MOSFETs are designed to turn on in the specific sequence depending on the Hall sensor signal. Meanwhile, rotor speed is adjusted through controlling six MOSFETs and PWM duty cycle. To develop the maximal braking capability, the phase current loop of the back-EMF needs to be utilized. When there is no driving current entering the motor stator and the motor remains inertial rotation, the motor will become a power generator with the current generated in the opposite direction to the driving current. Fig. 1 shows back-EMF of three phases denoted as ݄ܲܽ‫݁ݏ‬஺ ሺ‫ݐ‬ሻ, ݄ܲܽ‫݁ݏ‬஻ ሺ‫ݐ‬ሻ and ݄ܲܽ‫݁ݏ‬஼ ሺ‫ݐ‬ሻ. Since ܶ ൌ ʹɎ, ‫ݐ‬ଶ െ ‫ݐ‬ଵ ൌ ͳȀ͸ܶ , integral of the phase-to-phase back-EMF difference and armature current refers to the braking energy which can be used to stop the rotator. For example, the stopping power can be obtained as 1 16 T ( PhaseC (t ) − PhaseB (t )) I (t )dt . (1) T ³0

Fig. 1. Hall sensor signals and Back-EMF.

III. A.

DESCRIPTION OF CONTROL STRATEGY

Control process Fig. 2 illustrates the structure of the integrated driving and braking control system. First, when the motor works, the driving process is controlled by a PID controller.

887

The speed command is obtained from the throttle. Rotor speed can be detected and calculated from the Hall sensing signals. When the motor is switched to the braking mode, the MCU calculates the speed of the front wheel and rear wheel to estimate the slip ratio. Software

ŕũųŰŵŵŭŦ

v f (T )

Ĭ

e (T )

Į

őŊŅ

u (T )

Slip(T )

Hardware

D(T )

ŔŪŹĮŴŵŦűġŤŰŮŮŶŵŢŵŪŰů ħġőŘŎġŅŶŵźġńźŤŭŦ

v front (T )

v front (T )

œŰŵŰų őŰŴŪŵŪŰů

ŃųŢŬŦ ŔŸŪŵŤũ

SW

ŃųŢŬŦ ŎŰťŦ

v (T ) v front (T )

Į Ĭ

eslip (T )

Slip (T ) =

ŉŢŭŭ ŔŦůŴŰų

eslip (T )

v front (T )

⋅100%

Slip (T ) Fuzzy

Output command

Rule

v(T )

Hall C B A signal

ŔűŦŦť Braking Signal =1

v front (T )

Sector Output command

eslip (T )

d dt

Δslip(T )

ĴĮőũŢŴŦ ŊůŷŦųŵŦų

v (T )

ŇŶŻŻź ŔŭŪűġœŢŵŪŰ ńŰůŵųŰŭŭŦų

By considering the slip ratio and the change rate of slip ratio, one may implement a fuzzy logic controller to adjust the slip ratio. The fuzzy control rule base is established by considering the actual braking situation and observation form the real-world experiments.

Speed Calculation

D (T ) PWM Duty Cycle

ωmotor (T )

Fig. 3. Structure of the fuzzy slip ratio control.

ŃōŅńŉŎ

ω motor (T )

E.

ŇųŰůŵġŔűŦŦť

ŉŢŭŭ ŔŦůŴŰų

ω front (T )

Analysis of adding ultra-capacitor and braking torque controlled by PWM inverter

ŇųŰůŵġŘũŦŦŭ PWM Free Wheel

A+

B+

C+

PWM ON A+C-

PWM OFF

Hall:001

Fig. 2. Control system configuration.

Driver design The closed-loop control scheme for driving the brushless DC hub motor (BLDCHM) involves the difference between the actual and required speeds which is input to a PID controller. The controller adjusts the duty cycle of the PWM pulses that correspond to voltage amplitude to maintain the desired speed. The PID control command in terms of z-domain representation is given by

VC

B.

ª ( K p + K i + K d ) − ( K p + 2 K d ) z −1 + K d z −2 º uc ( z ) = « » e ( z ) (2) 1 − z −1 «¬ »¼

where ‫ܭ‬௣ , ‫ܭ‬௜ and ‫ܭ‬ௗ are the proportional, integral and derivative gains, respectively, and e( z ) denotes the speed error. ABS Control Design An ABS is a safety mechanism that allows the wheels on a motor vehicle to maintain tractive contact with the road surface according to driver inputs while braking, aimed at preventing the wheels from locking up and avoiding uncontrolled skidding. The key factor for implementing ABS control is slip ratio defined as follows

A-

v front (T ) − v(T ) v front (T )

t + Ts

³

where ࣰ௙௥௢௡௧ and ࣰ refer, respectively, to the vehicle and driving wheel tangent speed. For the EV considered in this research, the BLDCHM is installed at the rear wheel. The vehicle speed ࣰ௙௥௢௡௧ is calculated with the conversion gain ‫ܭ‬௙௥௢௡௧ times Hall signals obtained from the front wheel. D.

Control of slip ratio The fuzzy slip ratio control structure is shown in Fig. 3. When braking, the fuzzy slip ratio controller controls the slip ratio of the vehicle, and adjusts the PWM duty cycle to make the slip ratio fall into a safety range for the best tire-road surface adhesion to get the maximal braking torque.

888

vL ( t ) dt = D (Vdc − Vac − 2 IR ) − D ′ (Vac + 2 IR ) = 0

(4)

DVdc − Vac ,Vac < 0 2R

(5)

t

I=

where ܸௗ௖ is voltage of capacitor, D+D’=1, with D being the PWM duty cycle. The deduction of green line current flow is t + Ts

³ v (t ) dt = − D (V L

(3)

ON C-

L R Va=Vemf Vc=Vemf R L

One of six current flows corresponds to the current loop passing through the switches A+ and C- is shown in Fig. 4. The black line is the current flow of PWM duty cycle at the ON status, the red line is the current flow of PWM duty cycle at the OFF status. The green line is the inside circumfluence current which has nothing to do with PWM duty cycle. In this phase, the capacitor only discharges. The volt-second balance of black and red currents can be derived as follows

t

⋅100%

B-

Vb=Vemf R L

ON&OFF

Fig. 4. Current flow of the A+C- loop

C.

Slip (T ) =

Motor

bc

+ 2 I green R ) − D ′ (Vbc + 2 I green R ) = 0

I green =

−Vbc , Vbc < 0 2R

(6)

(7)

where ‫ܫ‬௚௥௘௘௡ relates to the value of back-EMF, and the braking current and torque are given, respectively, by I AC = I + I green =

DVdc - Vac Vbc + 2R 2R

§ DVdc - Vac Vbc · Tbrake = KT I AC = KT ¨ + 2R 2 R ¸¹ ©

(8) (9)

Thus, the larger PWM duty cycle represents that more current will flow into the motor, and the capacitor will discharge more.

2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA)

PWM

A+

B+

C+

PWM ON A+B-

PWM OFF

Hall:011

VC A-

ON

B-

Vb=Vemf R L

C-

ON&OFF


I green =

t +Ts

³

vL ( t ) dt = D (Vca − 2 IR ) + D′ (Vca − Vdc − 2 IR ) = 0

t

I=

( D − 1)Vdc + Vca , Vca > Vdc > 0 2R

³

(11)

vL ( t ) dt = D (Vcb − Vdc − 2 IR ) + D ′ (Vcb − Vdc − 2 IR ) = 0

t

I=

Vcb − Vdc , Vcb > Vdc > 0 2R

(12)

· § Vca − Vdc · ¸ + D′ ¨ 2 R ¸ ¹ © ¹

§ V −V · § V −V · I dc = D ¨ dc ab ¸ + D′ ¨ ca dc ¸ R 2 © ¹ © 2R ¹

PWM

A+

PWM

A+

B+

C+

Hall:011

VC A-

ON

B-

C-

Vb=Vemf R L

ON&OFF


The current flow of state 2 is shown in Fig. 6. The capacitor will charge and discharge. We discuss the current flow by two parts. The black and red line are the current flow discharged from capacitor. The green and bright green current flows are the PWM at ON status and OFF status, respectively. The volt-second balance of black and red line current can be derived from

³ t

PWM ON A+B-

C+

VC

vL ( t ) dt = D (Vdc − Vab − 2 IR ) − D ′ (Vab + 2 IR ) = 0

ON

C-

B-

(15)

PWM OFF

Vb=Vemf R L

L R Va=Vemf Vc=Vemf RL

Fig. 7. Current flow of state 3

Finally, the capacitor only discharges, the current flow of state 3 is shown in Fig. 7. The equation of volt-second balance is given by t +Ts

³ v ( t ) dt = D (V

dc

− Vab − 2 IR ) − D ′ (Vab + 2 IR ) = 0

(20)

t

I =

DVdc − Vab 2R

(21)

where the condition is

Vdc > Vca > 0 and Vab < Vcb < 0

Fig. 6. Current flow of state 2.

t + Ts

B+

Hall:011

L

PWM OFF

(19)

To charge the capacitor, the condition of ܸ௖௔ ൐ ܸௗ௖ ൐ Ͳ and ܸ௔௖ ൏ Ͳ must be satisfied. Hence, the bigger D represents that the more quickly the voltage of capacitor dissipates, and the less charge time will be. On the contrary, the greater D’ represents that the capacitor will charge more. In addition, the larger D’ and the smaller D refer to that the armature current reduces. Thus, the stopping torque will not be large.

A-

PWM OFF

(18)

For the capacitor, the Aampere-second balance is obtained

(14)

where Vca > Vcb > Vdc > 0 . This implies that the larger D′ represents that the capacitor will be charged more. In addition, the larger D′ and the smaller D represent that the armature current will be reduced so that the stopping torque will not be large enough. On the other hand, the larger D implies that the capacitor will be discharging with the shorter time. PWM ON A+B-

( D − 1)Vdc + Vca , Vca ≈≥ Vdc > 0, 2R

(17)

as

(13)

For the capacitor, the Ampere-second balance is given by § V − Vdc I dc = D ¨ cb © 2R

I=

(10)

The green line current is determined as follows t +Ts

vL ( t ) dt = D (Vca − 2 IR ) + D ′ (Vca − Vdc − 2 IR ) = 0

t

There are three states to be discussed when the current passes through the switches A+ and B-. The current flow of state 1 is shown in Fig. 5. The capacitor charges only. The current flow will be discussed by two parts. The volt-second balance of the black and red currents can be derived from

³

(16)

The green line and bright green current are determined as follows

Fig. 5. Current flow of state 1.

t + Ts

DVdc − Vab ,Vab < 0 2R

(22)

According to the above three states, for current charge of the capacitor in states 1 and 2, the lower PWM duty cycle at ON status, the more capacitor charge current will be. For the armature current, the larger PWM duty cycle at ON status, the more armature current and capacitor discharge current will be. In state 3, the capacitor only discharges; if the capacitor is not full, the discharge current and the armature current will be lower. Therefore, the reverse magnetic field will be weak, and the braking torque will not be large enough. We aim at finding the shortest braking time with the maximal braking torque. Different PWM duty cycles for the braking test, including 10%, 25%, 50%, 75% and 100%, have


889

been tested. See the final results of braking time summarized in Fig. 8.

B. Integrated braking and driving control system The integrated driving/braking circuit consists of a MCU (dsPIC30F3011), three phase inverters and an isolation circuit for gate driver. The braking control unit detects the motor speed by Hall sensing signals. Simultaneously, the MCU calculates an appropriate duty cycle based on the PWM technique to brake the BLDCHM by creating sequentially virtual short-circuit to the regenerative current loops. V. EXPERIMENTAL RESULTS

Fig. 8. Braking time vs. duty cycle.

The experiments show that the maximal raking torque occurs close to the 25% duty cycle. With regard to the above results for PWM duty cycle adjustment, when the slip ratio is gradually close to the boundary of the safety value, the output value of the PWM duty cycle is adjusted by the fuzzy logic rule to control the braking torque so that the slip ratio can be controlled within the safety range. IV.

ABS has long been adopted in traditional motor vehicles to prevent wheel from skidding during emergency brake. However, the effect is hard to be maintained on the iced surface. Our design has considered the following combination of road surfaces, the smooth and rough road under dry, wet, and waterlogging, respectively. Close up of the road surfaces are shown in Fig. 10 (a) and (b). The results of the road test and slip ratio control are collected, and the stopping time of the mechanic braking is compared to verify the slip ratio control and performance of ABS of our proposed design.

IMPLEMENTATION OF THE INTEGRATED DESIGN

A. Hardware implementation The electric scooter displayed in Fig. 9 is used to verify the proposed method. It is equipped with a brushless DC hub motor with 46 poles and a 120-degree Hall effect sensor at the rear wheel. Specifications of the motor are listed in Table I.

(a) Rough road surface (b) Smooth road surface Fig. 10. Close up of the road surface.

A. Dry and rough road surface Most roads are with dry and rough surfaces, which are considered first. The experiment conducted on that kind of road results yields the scooter speed and the slip ratio shown as in Fig. 11. The experiment result shows that the slip ratio remains less than 0.3, and the braking time is about 2.12 second.

Fig. 9. The experimental electric scooter. TABLE I Specifications of the BLDCHM Item Value Motor type brushless DC hub motor number of poles 46 poles Power 1kW Maximum 500 RPM Voltage 48V Rotor sensor type 120 degree Hall sensor Back-EMF type sine wave Fig. 11. Dry-rough road surface.

890


B. Dry and smooth road surface For the dry and smooth road surface, such as the white line on the road, the result of braking is shown in Fig. 12. On the dry and smooth road surface, the slip ratio exhibits a larger value. It may be up to 0.4, for the case, the braking time is 2.22 second.

Fig. 15. Puddles-rough surface.

F. Puddles and smooth road surface Fig. 16 shows the result on the puddles-smooth road surface. The braking time is 2.19 second. Fig. 12. Dry-smooth road surface.

C. Wet and rough road surface The test scenario for examining the performance of ABS on the wet pavement gives the braking time 2.11 second, see the result shown in Fig. 13.

Fig. 16. Puddles-smooth surface.

Fig. 13. Wet-rough surface.

G. The previous method On the flat ground, the braking times are very close, although the method proposed in this paper is tested on different road surfaces. The braking time is about 2.1 second for all cases, which is less than other two methods in large. On the road with puddles and the smooth surface, the braking time remain almost the same.

D. Wet and smooth road surface Fig. 14 shows the result collected in a wet and smooth road surface. The braking time is 2.14 second.

Fig. 17. Three lower arm short-circuit. TABLE II Experimental results under different conditions

Fig. 14. Wet-smooth surface.

E. Puddles and rough road surface Considering the road with puddles at the first section and then with a rough surface, the slip ratio is about 0.4 at the beginning and declines to 0.2 later on. For that case, the braking time is 2.21 second, which increases a little bit. The result is shown in Fig. 15.


891

The results are summarized in Table II. Compared with the short circuit braking, the braking time of the three low-arm braking method is slightly longer. As it can be seen from Fig. 17, when the vehicle with initial speed, the braking performance is slightly worsen. VI. CONCLUSION The purpose of this paper is to present a novel anti-lock electromagnetic brake for EVs that brakes the driving wheel with back-EMF on the motor to reverse the magnetic flux for enhancement of braking effect. The experimental results presented in this paper have shown significant improvement to the usual short circuit braking design with the shorter braking distance and stopping time. With the help of fuzzy slip ratio control, the braking torque is controlled to make the tyre slip ratio fall into the range for the optimal road face adhesion. This implements the function of ABS design as those of the traditional gas powered vehicles. According to the experimental results of the three lower arms, if the braking torque is insufficient, the slip ratio will not be within 0.2~0.3-the value to generate the maximal braking torque area. The three lower arms braking is a kind of braking methods that the less the vehicle speed is, the smaller the braking torque will be. Therefore, it is not suitable for high-speed braking which can better show the advantage of the proposed method here.

[11] S. N. Meitei, A. P. Kharghoria, U. K. Chetia, and S. Deka, "Regenerative braking along with ABS system in hybrid vehicles," International Conference on Computation of Power, Energy Information and Commuincation (ICCPEIC), pp. 317-320, Aug. 2016. [12] C. T. Huang, "An efficient energy recharging controller for electrical scooters," National Chung Hsing University, Master Thesis, pp. 1-65, Jun. 2009. [13] C. L. Lin, M. Y. Yang, E. P. Chen, Y. C. Chen and W. C. Yu "Novel antilock braking control system for electric vehicles," IET Journal of Engineering, 2017, under review. [14] C. L. Lin and W. C. Lin, “ABS control design for two-wheel drive electric vehicles,” International Conference on Mechanic Automation and Control Engineering, pp. 1011-1014, Inner Mongolia, 2011. [15] C. H. Tu, "Electromagnetic ABS for electrical vehicles using ultra-capacitor", National Chung Hsing University, Master Thesis, 2014. [16] W. C. Lin, C. L. Lin, P. M. Hsu, and M. T. Wu, "Realization of anti-lock braking strategy for electric scooters," IEEE Transactions on Industrial Electronics, vol. 61, pp. 2826-2833, Jun. 2014. [17] E. P. Chen, C. H. Tu, W. Z. Yu, C. L. Lin and Y. C. Chen, "New design of integrated driving and braking control system for electric vehicles," CACS International Automatic Control Conference, Kenting, 2017.

ACKNOWLEDGMENT This research was supported by Ministry of Science and Technology, Taiwan, R.O.C under the grant MOST 104-2221-E-005-093-MY2. REFERENCES J. R. Jan, "An experimental study of a DC brushless motor driving system," Da-yeh University, Master thesis, 2003. [2] Z. N. Shi, "Development of motor drive for motor," National Central University, Master Thesis, 2005. [3] K.S. Su, C.J. Lin, J.H. Lai and W.X. Wang, "Optimization design and driving circuit of brushless DC hub motor," Journal of Engineering Science and Education, vol. 11, pp. 333-347, 2014. [4] M. S. Antony and R. P. Raj, "Four quadrant operation of vector control of PMSM with dynamic braking," International Conference on Control Communication & Computing India (ICCC), pp. 161-164, Nov. 2015. [5] T. F. Lee, "The research of switched wheel motor windings applied to regenerative braking system," National Taipei University of Technology, Master thesis, Jul. 2007. [6] M. S. Antony and R. P. Raj, "Four quadrant operation of vector control of PMSM with dynamic braking," International Conference on Control Communication & Computing India (ICCC), pp. 161-164, Nov. 2015. [7] H. L. Jhou, "A novel method of electric braking with energy recovery for electric vehicles," National Central University, Master thesis, pp. 1-67, 2008. [8] J. Y. Su, "Study of regenerative braking for electric vehicle," Master thesis, Vehicle Engineering, National Pingtung University of Science and Technology, 2014. [9] X. Gong, S. Chang, L. Jiang and X. Li, "Research on regenerative brake technology of electric vehicle based on direct-drive electric-hydraulic brake system," Intional Journal of Vehicle Design, vol. 70, pp. 1-28, 2016. [10] O. C. Kivanc, O. Ustun, G. Tosun, and R. N. Tuncay, "On regenerative braking capability of BLDC motor," Annual Conference of the IEEE Industrial Electronics Society, pp. 1710-1715, Dec. 2016. [1]

892
