On the representation of sensor faults in fault detection filters

Presents an extension of the well known Beard-Jones detection filter that permits the isolation of sensor faults in a dynamic system to a fixed direction. The method is based on augmenting the system equations by an auxiliary state to represent the dynamic behavior of the sensor fault, and in effect converts the sensor fault into the same form as an actuator fault. The only condition required is observability of the original system. This method can be applied to both open-loop and closed-loop systems, and this is an advantage over traditional design methods