Abstract
The paper describes a theoretical investigation of failure detection using a new type of filter. The latter, called a residual difference generation detection filter, is based upon using (a) a main detection filter and (b) a faulty parameter modifier. Both of these use a reference model of the system; this is a mathematical model of the system without faults. In (a) the reference model is used in parallel with the real system in order to generate residuals (functions of differences between the real system and the reference model); in (b) the reference model is used in parallel with another mathematical model of the system which contains a known fault which again generates residuals that are a function of the known fault. By analysing these residuals while changing the known fault in (b) it is possible to detect which fault is occurring. The technique is applied to faults in a servo-motor, faults being assumed to occur singly. Tests were carried out using a mathematical model of the real system which incorporated an ‘unknown’ fault.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have