Research Of Model Based Diagnosis For Vehicle Active Suspension Control Systems

With people’s increasing demand for safety and comfort of vehicles,the suspension structure has been continuously upgraded and improved to meet the performance requirements of modern automobile.Active suspension system has received wide attention and application for its great comfort.However,due to its complex structure,the probability and frequency of fault occurrence are much higher than that of traditional suspension systems.In order to improve the safety of active suspension system,it is necessary to use the fault detection and diagnosis method to monitor the active suspension in real time.In this paper,the observer-based fault diagnosis method is used to detect and diagnose the faults of the active suspension system actuators such as stuck and loss of effectiveness;the parity space method is adopted to detect sensor faults of the active suspension system;the feasibility of above methods for fault detection and diagnosis in active suspension is verified by simulation.And main research work is summarized as follows:First,the significance of fault detection and diagnosis of active suspension system is introduced and the history of fault diagnosis technology is briefly reviewed.The related concepts and basic problems of fault diagnosis are described.At the same time,modern fault detection and diagnosis methods are introduced and classified in detail and its application in vehicles is introduced.Secondly,the active suspension system is briefly introduced.Based on the dynamic theory,the mathematical model of the active suspension system is established;lock in place,loss of effectiveness and no-output failure of the active suspension actuator are modeled;the mathematical model of sensor faults is also established.Thirdly,a robust estimation method based on unknown input observer is proposed to detect the actuator fault of active suspension system with unknown input signal,which realizes the decoupling of unknown input.And the best matching method is used for fault detection and fault isolation.According to Lyapunov stability design adaptive rate,the size of fault is estimated by adaptive method.To improve the real-time performance of the diagnosis process,an improved fault diagnosis algorithm is designed to ensure the real-time performance of the diagnosis system.Fourth,fault diagnosis algorithm for the active suspension actuator with model uncertainty is carried out.The algorithm is based on Luenberger adaptive observer and applies the H_∞performance index to the adaptive algorithm.And observer with robust residual is designed to estimate the size of the fault.The model uncertainty problem is solved by the method that satisfies the H_∞ performance index.Then the H_∞ problem is solved by the linear inequality to obtain the observer parameters.Fifth,The parity space method is used to detect and isolate sensors fault of the active suspension system sensors.The traditional direct redundant parity space method is introduced.Based on the direct redundancy method,the software redundant parity space method is deduced,and the software redundancy method is used to detect the faults of active suspension system sensors.Taking into account the effects of system noise,the algorithm is designed to reduce the effect of system noise on the detection results for unknown input disturbances.