A robust fault diagnosis system of packaging process based on multisensor and probabilistic neural networks

In this paper, we put forward a fault diagnosis system based on multisensor and probabilistic neural networks (PNN). The multisensor is employed as an effective direct abstracting tool for the frequency segments features of the field signal. The fault signals are dealt with by using probabilistic neural network estimating models. Simulation results show that the proposed method is featured by swiftness, accuracy and ease of practical application.