FAULT DETECTION IN FLUID SYSTEMS: AN INTERACTIVE REAL-TIME APPROACH

Model-based fault detection is typically realised by evaluating the possible deviation between expected and real behaviour of a system. In the cases that this comparison is performed under dynamic conditions during the normal operation of an industrial system a special kind of models is needed for this interaction. In this paper an implementation model is presented where simulation results and real-time data acquired from an actual system are effectively compared on-line in a suitable environment and the residuals are evaluated for the fault detection process. The experimental results show that the proposed method is reliable and can be applied to real world systems.