EVS8 KINTEX, Korea, May 3-6, 15 Fault-tolerant Control System for EMB Equipped In-wheel Motor Vehicle Seungki Kim 1, Kyungsik Shin 1, Kunsoo Huh 1 Department of Automotive Engineering, Hanyang University, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, Korea Department of Automotive Engineering, Hanyang University, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, Korea, Tel.: +8---437, E-mail: khuh@hanyang.ac.kr (corresponding author)
Contents Introduction Brake Actuator Modelling Vehicle Dynamics Controller Design Braking Force Distribution Logic Simulation Conclusion
Introduction Motivation Popularization of EVs Increase in interest in fuel economy and regulation of the environment Replace the ICE(Internal Combustion Engine) with the Electric Motor In-Wheel Motor Vehicle is being studied Demand for the New Type of Brake System The conventional hydraulic brake system equips the vacuum booster EVs need new type of brake system because they have no ICE and vacuum booster. EMB(Electro-Mechanical Brake) System Using the Electric Motor Fast response and Less Weight No environmental pollutant Re-generative Brake System In EVs, the re-generative braking by drive motor is available In In-wheel Motor Vehicle, independent braking of each wheel is available 3
Introduction Related Work: Component Level Fail-safe Approach Fault-tolerant control of EMB systems (Ki et al., SAE Int. J., 1) The fail-safe system utilizing the estimated signal when sensor failure was occurred Sensor type: Current, Position, and Clamping force etc. This research was focused on the component level fail-safe didn t consider the system level fail-safe <Block diagram of EMB control system> <Flowchart of sensor fault-tolerant control logic> 4
Introduction Related Work: System Level Fail-safe Approach Control of brake- and steer-by-wire during brake actuator failure (Hac et al., SAE Technical Paper, 6) Development of a fail-safe control strategy based on evaluation scenarios for an FCEV electronic brake system(jeon et al., IJAT, 1) Simplified fault-tolerant control algorithm is used in these researches <Algorithm flow chart for BBW vehicle> <Fail-safe control strategy for one EMB failure> 5
Introduction Related Work: System Level Fail-safe Approach (Cont d) Fault-tolerant control with active fault diagnosis for four-wheel independently driven electric ground vehicles (Wang et al., IEEE Transaction on Vehicular technology, 1) The active fault diagnosis is proposed to explicitly isolate the faulty EMB wheel and to estimate control gain of the faulty wheel The adaptive control scheme is used for calculating the desired braking torque at each EMB <Block diagram of the proposed control system with active fault diagnosis> <Control gain estimation on the wheel> 6
Introduction The Purpose of Fault-Tolerant System Minimize the Effect of Faulty EMBs Compensation of Total Braking Force Increase the Braking Torques from No fault occurred EMB Prevention of the Partial Brake Satisfy the Driver s Braking and Steering Command Enhancement of the Vehicle Stability Consider the Limits of braking Road-Tire Friction Limits Magnitude of generated tire force is decided by tire-road friction coefficient and vertical load Therefore, no matter how large clamping force is produced, there is limitation of generated braking force on each wheel Brake Actuator Performance Limits Due to the limitations of the Actuator, the maximum braking force is determined. : The Fault-tolerant system should be designed to consider the limits! 7
Fault-tolerant System Architecture System Architecture Vehicle Dynamics Controller (Sliding Mode Control) The required total braking force and moment is generated to fulfill the intend of the driver Braking Force Distribution Logic (Optimization) Each braking force of EMB is calculated to track the desired braking force and moment at the same time satisfy the constraints Sensors Constraints Calculation Driver Inputs Vehicle Dynamics Controller Desired Force Desired Moment Constraints Braking Force Distribution Logic EMB Braking Command Rear-Right EMB Rear-Left EMB Front-Right EMB Front-Left EMB Available Re-generative Braking Torque EMB Status Monitoring Battery SOC Controller Re-generative Braking Command 8
EMB(Electro-Mechanical Brake) Model Experiment & Fitting Input Current Vs. Clamping Force The EMB has linearity between current input and clamping force F K u i_ clp i i The Braking force generated from EMB can be modelled as: F F F i_ clp i_ EMB K u i i : EMB Clamping Force : EMB Braking Force : Friction Coefficient b eff rf rk b i_ clp b i i_ EMB i reff reff : EMB Linear Gain : Input Current rb : Effective Brake Rotor Radius r : Effective Wheel Radius u Clamping Force [N] 3.5 x 14 3.5 1.5 1.5 EMB Hils Data 5 1 15 5 Current [A] 1A A 3A 4A 5A 6A 7A 8A 9A Linear Fitting 9
Re-generative Brake Model The Strategy of Regenerative Braking The maximum available regen. torque is determined depending on the speed of the vehicle The maximum available regen. torque is depending on the speed of driving motor The regenerative torque varies depending on: The SOC (State Of Charge) for managing the battery The fault status of the EMB The regen. torque is maximized when the fault on its EMB is occurred Re-generative Braking Torque [Nm] 5 15 1 5 Maximum Re-generative Braking Torque Max. Regenerative Braking Torque % Regenerative Braking Torque 5% Regenerative Braking Torque 4 6 8 1 1 14 16 18 Vehicle Speed [kph] 1
Vehicle Dynamics Controller Design Driver Model Desired Longitudinal Speed V V a dt x, desired x, desired Desired Yaw Rate Vx sw desired mv x ( lrcr l fc f ) GR lf lr C C ( l l ) f r f r Sliding Mode Controller Design Newton s Law mv F F I v x x x, disturbance M z cg Control Law F m V sat S F x, desired v x, desired 1 1 x, disturbance M I sat S cg, desired z desired Sliding Surface S V V 1 x S d xd S 1 V x V xd 1sign( S1) S sign( S ) d S i, if Si i S i i sat i S i sign, if Si i i 11
Braking Force Limit The limitations on the braking force The normal force acting on each wheel considering the acceleration Mb Mh Mh NLF g ax f ay L L tw Mb Mh Mh NRF g ax f ay L L tw Ma Mh Mh NLR g ax r ay L L tw Ma Mh Mh NRR g ax r ay L L t Limits on the road friction coefficient F N i, RoadLimit tire i w N F l f L h lr N R Limits on the EMB Actuator F rk b i iemblimit, ui,max reff F x The maximum braking force of each wheel F min F F, F i,max i, EMB Limit i,regen i, Road Limit F y 1
-D.O.F Vehicle Modeling Ratio of the Maximum Braking Force Maximize the stability of vehicle and margin of the actuator while braking F F i i i,max Total : 5N Total : 5N The Braking Force and Moments Acting on the Vehicle Summations of the braking forces on left and right side F R1 F F F F F F L1 FL F R w x_ cg L1 L R1 R w M F F F F cg L1 L R1 R 1,max,max 1,max,max F F F F F M x_ cg L L1,max L,max R R1,max R,max cg F F F F w L L L R R R 13
Optimization Problem Object Function Conditions on the required braking force F F F F F x, desired L L1,max L,max R R1,max R,max F L1 F R1 Conditions on the required moments w M F F F F cg, desired L L1,max L,max R R1,max R,max FL F R w Formulation of the object function J W F F F F F x F x, desired L L1,max L,max R R1,max R,max w W M F F F F cg M cg, desired L L1,max L,max R R1,max R,max Constraints The ratio of the braking is bounded from to 1 1 L 1 R g1: L g: 1 L g3: R g4: 1 R 14
Convex Optimization Formulation of the Lagrange Function KKT(Karush-Kuhn-Tucker) Necessary Condition Necessary Condition Sufficient Condition H L J u g s L i L ui L si J H WFWM w FL1,max FL,max FR1,max FR,max det 4 x cg Positive Definite i i i where J:Object Function g i :Constraints u i :Lagrange Multiplier s :Slack Variable The solution set, which not only satisfies the constraints but also minimizes the object function, is exist! Convex Optimization Problem i 15
Simulation Configuration Simulation Tools MATLAB/Simulink & CarSim Simulation Scenarios Friction coefficient of the road:.85 Steering input: Double Lane Change Speed range: 1 km/h ~ km/h 1.33m.5544m.64m Vehicle Parameters M 187. Curb weight M f 114. M r 73. Effective radius of the wheel r w.38 Wheel base L.64 Dist. of F wheel to CG point L f 1.33 Dist. of R wheel to CG point L r 1.677 Height of the CG point h.5544 Effective radius of front disk radius_eff_front.1114 Effective radius of rear disk radius_eff_rear.113 Front brake pad friction coefficient U_pad_front.34 Rear brake pad friction coefficient U_pad_rear.34.38m 1.585m Front :.1114m Rear :.113m :.34 16
Simulation Results No faults on any EMB (.3g Command w/o Re-generative Brake) Algorithm based on the Look-up table vs. proposed method 1 1 Vehicle Velocity Desired Vx Vehicle Vx 1 1 Vehicle Velocity Desired Vx Vehicle Vx 8 8 velocity [kph] 6 4 velocity [kph] 6 4-5 1 15-5 1 15 17
Simulation Results No faults on any EMB (.3g Command w/o Re-generative Brake) Algorithm based on the Look-up table 4 3 Vehicle YawRate Desired Yaw Rate Vehicle Yaw Rate 5 EMB Braking Force FxL1 FxL FxR1 FxR yawrate [deg/s] 1-1 braking force [N] 15 1 - -3 5-4 5 1 15 Algorithm based on the proposed method 15 1 Vehicle YawRate Desired Yaw Rate Vehicle Yaw Rate 5 1 15 3 5 EMB Braking Force FxL1 FxL FxR1 FxR yawrate [deg/s] 5-5 braking force [N] 15 1-1 5-15 5 1 15 5 1 15 18
Simulation Results Fault occurred on the Front-Left EMB (.3g Command w/o Re-generative Brake) Algorithm based on the Look-up table vs. proposed method 1 1 Vehicle Velocity Desired Vx Vehicle Vx 1 1 Vehicle Velocity Desired Vx Vehicle Vx 8 8 velocity [kph] 6 velocity [kph] 6 4 4 5 1 15-5 1 15 19
Simulation Results Fault occurred on the Front-Left EMB (.3g Command w/o Re-generative Brake) Algorithm based on the Look-up table 4 3 Vehicle YawRate Desired Yaw Rate Vehicle Yaw Rate 5 EMB Braking Force FxL1 FxL FxR1 FxR yawrate [deg/s] 1 braking force [N] 15 1-1 - 5-3 5 1 15 Algorithm based on the proposed method 15 1 Vehicle YawRate Desired Yaw Rate Vehicle Yaw Rate 5 1 15 4 35 3 EMB Braking Force FxL1 FxL FxR1 FxR yawrate [deg/s] 5-5 -1 braking force [N] 5 15 1 5-15 5 1 15 5 1 15
Simulation Results Fault occurred on the Front-Left EMB (1.g Command w/ Re-generative Brake) Braking performance: The braking command of 1.g is not satisfied (.5g approx.) Stability performance: The steering stability is satisfied while on braking 1 1 8 Vehicle Velocity Desired Vx Vehicle Vx 6 5 4 EMB Braking Force FxL1 FxL FxR1 FxR velocity [kph] 6 4 braking force [N] 3 1-5 1 15-1 5 1 15 15 1 Vehicle YawRate Desired Yaw Rate Vehicle Yaw Rate 5 Re-generative Braking Torque L1 (1%) L (3%) R1 (8%) R (3%) 5 yawrate [deg/s] -5 braking torque [N] 15 1-1 5-15 5 1 15 5 1 15 1
Conclusion Fault Tolerant Algorithm Design The fault tolerant algorithm is designed to minimize the effect of the faults Compensation of the lack of the total braking force Prevent loosing stability from the differential braking The braking force re-distribution logic is designed considering the limitation of braking force The limitations of the friction coefficient between the road and tire is considered The performance limitations on actuators are considered Validation Simulations with the commercial software, the MATLAB/Simulink and CarSim Effects of the faults on EMB are evaluated The proposed algorithm in this research is validated Future Work More detailed vehicle model will be evaluated to design the algorithm including the steering model Reasonable formulation of the object functions