Reconfiguration Using a Virtual Sensor

Abstract

This chapter develops the concept of a virtual sensor: when a sensor is at fault, an observer is used to calculate a replacement value. This approach goes back to the 1970s when it has been studied in the context of fault detection as “dedicated (and generalised) observer scheme” (see Schröder [2003] for a modern version of this idea). However, in the literature the observer has only been used in combination with state feedback, while a general dynamical controller will be considered here. It will be shown that the virtual sensor can be used as a reconfiguration block by translating the measurements from the faulty plant into the values the controller can handle. The stability of the reconfigured loop can be guaranteed as long as the faulty plant is detectable. The results found here are the basis for the derivation of the virtual actuator in the following chapter.