Model-Based Fault Diagnosis and Prognosis for Electric Power Steering Systems

Electric Power Steering (EPS) is an advanced steering system that consists of two subsystems: electrical and mechanical subsystems. EPS systems not only provide steering assist to drivers but they are also actuators for recently developed active safety features, such as lane keeping and lane changing assist. Failure of some component of the EPS system can lead to walk-home situations and increased warranty costs. Hence, for the improvement of reliability, safety, and efficiency of EPS systems, fault detection, diagnosis, and prognosis become increasingly important. This paper provides fault detection for EPS systems through model-based techniques using parameter estimation to determine the current electric parameters of the EPS motor. In addition, by monitoring the deviation of the self-aligning torque (SAT) estimated from two different methods, changes in EPS mechanical parameters can be detected. The progression of this deviation can be fed into a health state estimator which can give an indication of state of health and remaining useful life. Computer simulations as well as hardware-in-the-loop (HIL) experiments are provided to illustrate this method. Finally, for integrated system diagnosis and fault isolation, a fault signature table is constructed based on estimations of motor parameters, calculations of road SAT, and residuals of parity equations. This table can be used to detect and isolate considered electrical, mechanical, and sensor faults in the EPS system and simulation results are shown to verify the developed ideas.