Research On Fault Diagnosis And Fault Tolerant Control Of Automobile Semi-active Suspension Actuator

With the development of automotive industry and people's consumption upgrade,the performance of a car's ride comfort,handling stability,reliability,and safety is increasingly drawing attention.As a result,the controllable region of a car suspension system has been widened to meet such demands.As an example,semi-active suspension system was invented for this purpose with the help of sensors,as a main measurement channel of the suspension system,and a control module,which manipulates actuators,as a main channel for the control input,to implement control requirements.A risk is companied with such designs of semi-active suspensions.When the sensors and the actuators suffer from a fault,the control module,which is designed with a fault-free assumption,may malfunction so that the expected control performance will degrade and even fail to have effect on the suspension system completely.Relying on the National Natural Science Foundation Youth Fund Project ‘Research on damping control and adaptive switching of car active suspension systems based on intelligent road perception'',and the Joint Fund Project of Liaoning Science and Technology Department‘Adaptive control of car active suspension systems based on intelligent road perception',this thesis studied the fault diagnosis methods and fault tolerant control for a damping-adjustable semi-active suspension system,to solve the problem caused by the fault-free assumption on suspension parts.This research has the potential of improving the reliability and quality of suspension control and providing optimal methods for suspension design.Specifically,the research was carried out from the following aspects:(1)A dynamics suspension model,which includes a quarter vehicle model with two degrees-of-freedom(DOF)and a seven-DOF vehicle model for describing the vibration of the suspension,was developed.Based on the road roughness coefficient,a single-wheel model for random road inputs and a four-wheel model for correlated random inputs were built.Without the consideration of sensor faults,the fault models for actuator-gain,constant deviation,and stuck faults were devised.(2)Based on the classic Kalman filter algorithm,a Kalman state estimator was designed for the estimation of vehicle vibration states,including the vertical displacements and the vertical velocity of sprung and unsprung masses.Matlab/Simulink simulation was conducted to evaluate the estimation performance to establish a basis for the fault diagnosis of suspension actuators.(3)The Kalman state estimator were applied to diagnosing possible suspension faults based on the residual theory.A LQG controller were devised for the fault tolerant control through the compensation of suspension control forces.Matlab/Simulink simulation for the verification of the control performance were carried out.(4)A hardware-in-the-loop simulation environment was established for the verification and evaluation of the proposed fault diagnosis algorithm and the fault tolerant control algorithm.The simulation and experimental results show that the designed state estimator was able to accurately estimate the states of the suspension system and detect faults in time.The compensation control force was effective in resuming the functions of the fault suspension and eliminating the adverse effects,which improved the reliability of the active suspension control system.