Application Research of Artificial Intelligence in Fault Diagnosis
Tian Li · Journal of Liaoning University · 2012
According to the drawbacks and the shortages of the neural network,this paper analyzes the neural network and rough set,support vector machine,wavelet and particle swarm combination method.Combining rough set with neural network in fault diagnosis can eliminate redundant information,reduce the neural network's input layer node,simplify network structure,shorten training time through the rough set for reduction of knowledge;The composite fault diagnosis technology based on neural network and support vector machine not only can further improve the prediction accuracy of the individual models,but also make a fault diagnosis model always in the optimum recognition state;Particle swarm optimization neural network can speed up the network convergence speed and improve training accuracy;Wavelet neural network classify fault very well,and has high recognition accuracy.