Abstract
The paper presents a new method for fault detection in discrete event systems that was developed based on the classical diagnoser approach. The goal was to reduce the amount of computation that has to be done for isolating a fault as soon as possible. To do this, we have reduced the number of states for the process model, the system model and consequently, the number of states for the diagnoser. In order to demonstrate, the method is used and tested on a compressor station, in a comparative analysis with the classical method.