Fault Detection of Vehicle Suspensions

Abstract In this paper methods for fault detection and diagnosis of vehicle suspensions are presented. With parameter estimation and parity equations symptoms on the current process status can be extracted. By means of parameter estimation detailed symptoms can be generated, allowing an automatic distinction of different faults. For the automatic classification, neural networks were trained. Parity equations are well suited for a detection of sensor faults, although they do not offer the possibility for a distinction of different faults. All the presented results were obtained with measured data drawn from a test rig and a driving car. The proposed methods are suited for an on-line supervision of the car as well as for technical inspection e.g. in workshops.