An intelligent system for integrated predictive diagnosis

We present an automated system for integrated predicted diagnosis. This fault diagnosis method was tested on vibration data recorded from an aft main power transmission of a US Navy CH-46E helicopter. The fault diagnosis system is based on a neuro-fuzzy algorithm. First frequency domain analysis techniques were used to extract features from the vibration signals. These features were then clustered by a self organizing map neural network and identified by a backpropagation network. The decisions from different channels or sensors were fused using fuzzy logic techniques.