Model-based fault detection for time variant processes with application to an automotive actuator

Automobile actuators have to work very reliably. The application of hardware and software redundancy schemes is a well known technique for the improvement of safety and reliability. In extension to that, the use of model-based fault detection methods is advantageous for the early detection of incipient faults in the process or in the signals. In this paper, an approach is presented which combines parameter estimation with parity equations in order to cope with the time variant characteristic of technical processes. After a brief description of the theoretical basis the experimental application follows for the example of an automobile actuator which is driven by a DC motor