×

A fault detection method based on the fusion of computational intelligence approaches. (Chinese. English summary) Zbl 1240.68169

Summary: Various approaches, namely, rough set, genetic algorithm and neural network, are integrated to synthesize their merits for fault detection. According to the uncertainty and imperfection of the original data sample, rough set is used to pretreat the data, i.e., for the normalization of data, the discretization of continuous data and attribute reduction, in order to obtain a minimum fault feature subset. A genetic algorithm, having the ability of strong global search, is used to train the weights of a back propagation neural network. The minimum reduced subset is used as input for the trained network to construct the fault detection model which classifies the pretreated fault feature vectors under certain states to realize the fault detection. Experimental results on motor bearing show that the method is able to optimize the structure of the neural network and to improve the rate and precision of fault detection.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
68T20 Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.)
PDFBibTeX XMLCite
Full Text: DOI