Study And Develop On Engineer Aid-analysis System In Fault Diagnosis Based On Wavelet Neural Network

Most of the modern industry equipments are big or even huge systems that consisted of many parts, and the relationship among them are getting more compact and intricate. They should not be seen departed. However the traditional diagnosis that totally depends on people's experience can not fulfill the industrial whole system's diagnosis need. So the developing of advanced fault diagnosis that fit with modern industry is necessary.Because of its unique characteristics, computers are broadly used in lots of fields, especially in industry. The combination of fault diagnosis and computer changed the situation that can't put theories into applications. From the existed abundant experiences and lots of edge theories comes a lot of new usable technologies. They boosted fault diagnosis into the intelligent stage.Focused on large-scale rotate machine, this paper introduces some kinds of typical fault's mechanism, their expert knowledge and means to deal with them. It also expound the appliances of diagnosis methods on equipment state monitoring tech and fault diagnosis, Particularly the vibration diagnosis which is very typical and the signal processing method related.This paper introduces the mechanisms, functions, models of wavelet and neural network. By combining them, it deduced the mathematics model of three layer RBF neural wavelet network. Then built the model of intelligent fault diagnosis and introduced the function and model of each module in detail. With the actual experience and powerful tool of VC++6.0, a software was developed.The system can be used in simple fault recognition after learning the expert knowledge which was input in its demanded format.