A new fault diagnosis method to hydraulic system for die casting machine based on FARX and RBFNN models was put forward.The ARX model after fuzzy processing not only retained its very sensitive response to fault state pattern

but also achieved more accurate description of non-linear features.Firstly

several FARX models were established for target system fault to extract their fault vectors.Secondly

the vectors were served as training source data and the RBF network classifier was established.Then

fault patterns could be identified according to the output of the classifier.A simulation verified the accuracy and availability of the method.