What are methods for determining errors in a real system?
The aim of a control system is to force some of the system variables towards a desired state. In order to achieve this, one must be able to measure the actual state of the system and compare this value against the desired state. The difference between the desired and the actual measurement produces an error. This error is then used as an input to a controller whose output signal forces the system towards the desired state.
Errors then are typically obtained by determining the difference between a desired value and the actual or measured value. This idea works also for fault detection in systems where one knows the nominal response of a system to a known signal and an error is produced if the system starts to deviate from this response.