ABSTRACT
Chapters 8 and 9 introduced the Kalman filter and the neural network for fault isolation in gas turbine engines, respectively. Fuzzy expert systems are robust and are being increasingly used for diagnostics and other applications [95-98]. Recently, it has been proven that classical feed-forward neural networks of the type used in engine diagnostics can be approximated to an arbitrary degree of accuracy by a fuzzy logic system, without having to go through the laborious training process needed by a neural network [99]. In addition, fuzzy rules follow human language-based reasoning processes and are much easier to interpret and understand than neural networks that have a black box nature [100]. In this chapter, it is shown that fuzzy logic systems can be used for engine module fault isolation under high levels of uncertainty. In addition, fault isolation results from the fuzzy logic systems are compared with results from neural networks and Kalman filter methods. The application of fuzzy logic to gas turbine diagnostics, which is discussed in this chapter, was introduced by Ganguli [13, 101].
