Study about fractal neural network diagnosis method and application

In This paper, Fractal calculating dimension is firstly put forward. Combined Fractal theory with Neural network, A Fractal Neural network identification methods is built and applied to the state control and fault diagnosis of Mechanical equipment. This network is made of three layers construct: Input layer, hide layer and output layer. Input and output of standard samples are respectively Fractal calculating dimension of different period sampling and the unit matrix equal to sample numbers. Weight and threshold of network is rapidly and correctly computed by conjugate terraced optimization. Rolling bearing fault is perfectly identified by this diagnosis way.