Condition Diagnosis of Blower System Using Rough Sets and a Fuzzy Neural Network

摘要

This paper presents a condition diagnosis method for a blower system using the rough sets, and a fuzzy neural network to detect faults and distinguish fault types. In order to solve the ambiguous problem between the symptoms and the fault types, the diagnosis knowledge for the training of the neural network is acquired by using the rough sets. The fuzzy neural network realized by partially-linearized neural network (PNN), which can automatically distinguish the faults. The PNN can quickly converge when learning, and can quickly and high-accurately distinguish fault types on the basis of the probability distributions of the machine conditions when diagnosing. The non-dimensional symptom parameters are also defined in frequency domain, and those parameters are processed by rough sets to sensitively diagnose machinery conditions. Practical examples of the diagnosis for a blower system are shown in order to verify the efficiency of the method proposed in this paper.