Research Of Diagnosis Method For Vehicle Steer-by-wire System

Since Steer-by-wire(SBW) system eliminates the mechanical connection between the steering wheel and the front wheels, it fully overcomes the disadvantages of conventional steering system and provides the potential benefits of coordinative and fully integrated vehicle dynamic control. However, SBW includes many electrical components; SBW is less dependable than mechanical system. Fault diagnosis is essential to maintain the safety and dependability of SBW system. The dependable level of ECU is higher than sensors and DC motors, therefore, the diagnosis methods of sensors and DC motors are provided in detail here. As the analytical model of SBW is easy to obtain, model-based diagnosis method is adopted. For sensor diagnosis, state estimation is used to build analytical redundancy. By comparing the measuring signals and estimating signals, a residual vector is established. Different sensor fault appears in different direction of the residual vector, so the fault detection and isolation(FDI) of sensor can be achieved. The state estimation algorithm used here is an adaptive Kalman filter(AKF), which overcomes the disadvantages of conventional Kalman filter when applied in the parameter-variable SBW system. For motor diagnosis, AKF-based parameter estimation is used to watch over the motor parameters(such as resistance, inductance), once the parameters go beyond the normal range, the motor fault is detected. In the forth chapter of this thesis, the diagnosis model for SBW is built by Matlab/simulink, and simulation is applied to verify the diagnosis method provided in this thesis.The specific research of this thesis includes:1,The diagnosis method for sensorsThere are many sensors in SBW system: later acceleration sensor and yaw rate sensor used for reflecting the vehicle state, current sensor and angle sensor used for DC motor control, angle sensor used for perceiving driver's steering intention. Generally these sensors are composed of precise components. Working in bad environment, senor is inevitable to fail. Once sensor fault happens, part or all of system function will lose.A simple idea of sensor diagnosis is: building redundant sensor measurement values, then comparing redundant sensor measurement values and real sensor measurement values to establish a residual vector, through the logic threshold, FDI can be achieved. Because of cost problem, hardware redundancy is replaced by analytical redundancy. Analytical redundancy makes use of the potential redundant relation between concerned sensors and other sensors in SBW system. With a certain algorithm, concerned sensor value can be estimated. In SBW, front wheel angle is a very important control signal, so sensors for front wheel angle measuring must be diagnosed. Without adding another angle sensor, two redundant angle signals can be obtained from later acceleration sensor and yaw rate sensor through a certain filter algorithm.FDI method for angle sensor used for perceiving driver's intention is not discussed here. There is no mechanical connection between the steering wheel and the front wheels, therefore, there is no dynamic relation between the two parts. It is difficulty to introduce an independent analytical redundancy. The diagnosis of steering wheel angle sensor mainly depends on hardware redundancy. Two angle sensors perform mutual supervision, once there is significant difference between the two, sensor fault must happen.2,The diagnosis method for DC motorsThere are two DC motors in SBW system, the steering motor and road feeling motor. Road feeing motor receives the torque signal from the main controller and creates steering wheel return torque, providing driver with road feeling. If road feeling motor fails, driver will lose road feeling. Steering Motors receive main controller's command and achieve the intention of driver's steering. If steering motor fails, driver will loses steering.Real time estimation of motor parameter (resistance, inductance) is a method for motor diagnosis. If the motor is in good condition, the parameters are in a normal range. Once the parameters go beyond the normal range, the motor fault will be detected.3,SBW FDI system simulation by Matlab/simulinkUsing Matlab/simulink, SBW FDI system is established. FDI logic for sensors diagnosis are tested and verified, the simulation results demonstrate that sensors diagnosis method in this thesis is valid. Resistance estimation application in motor diagnosis is simulated. The validity of this method is also verifiedAbove research provides a good foundation for the SBW fault tolerant system.