Research on fault detection and diagnosis strategy based on AMT system behavior

Abstract

Fault detection and diagnosis strategy based on AMT (Automatic Mechanical Transmission) system trajectory is studied. Hybrid automata is used to build healthy AMT system model. Then fault definition of AMT system is analyzed from the point of system trajectory and fault diagnosability of AMT hybrid system is studied based on AMT system's fault model and fault threshold function. According to AMT system model trajectories and control cycle, AMT hybrid system's behavior is analyzed and classified into three types: steady-state system behavior, discrete event trigger behavior and physical system response behavior. At last AMT system's fault detection and diagnosis strategy is proposed based on behavior types. The fault diagnosis strategy with the diagnostic algorithm is transplanted to the real car platform and a large number of vehicle mileage prove the strategy's correctness and real-time.