ABSTRACT
In this paper, we discuss the structure of Self Generating Neuron-Fuzzy Model, a self generating algorithm and its application to machinery condition monitoring. First, we analyze the neuron-fuzzy model based on radial basis function (RBF). Then, we discuss the concrete algorithm of this model, in particular, we improve the algorithm of the model for the application of machinery condition monitoring. At last, we use the model to identify modes of bearing operating condition. The results are shown that the model is adaptive according to the specified model error with less number of RBFs than the other methods by which only coefficients of the RBFs are tuned. So the model satisfies the basic requirements of machinery condition monitoring, such as real-time property, success rate, sensitivity and robustness.
