The Research Of Gearbox Intelligent Diagnosis Technology Based On Demodulation Analysis And BP Neural Network

With the improvement of the complexity and the level of automation of the machinery and equipments,fault diagnosis for mechanical equipment is being more important significantly,and the choice of appropriate diagnostic methods for the accurate diagnosis is being more important. For knowledge representation and reasoning as a difficult point in the intelligent system which studied by a lot of people,the rules contain a comprehensive fault type is a question worth considering.Artificial neural network to imitate the human brain's physical structure,with its powerful ability of parallel computing and the association is very suitable for mechanical fault diagnosis,neural network technology of their parallel distributed processing of information in fault diagnosis is becoming superiority.In order to extract the early data of fault information,and intelligent fault diagnosis more effectively,a gearbox intelligent diagnosis study based on demodulation analysis and neural network is proposed in this paper.First of all,this paper introduce the importance of technology and the development and trend of machinery fault diagnosis at home and abroad,pointing out that the artificial neural network theory is applied to machinery fault diagnosis technology has great application value and development potential.At the same,also pointed out that China's current artificial neural network theory in the field of machinery fault diagnosis application is not perfect,is still in development stage.And the main research is the demodulation method of early fault featres extraction,based on neural network technology,intelligent diagnosis system is a improvement for existing intelligent diagnosis system of high-speed wire rod mill.Including the supplementary of the existing case library and rule base,laying a foundation for fault intelligent diagnosis;rules'supplementary through demodulation analysis and methods based on the EMD decompositon to extract fault feature to supplement the traditional spectral analysis.These diagnosis parameters of signal analysis methods can be used to learn and train as input vector for neural network.Neural network technology can rich gearbox function of intelligent diagnosis.On the orverall,CBR and RBR combined with neural network enriches intelligent diagnosis system and makes it more effective to carry out intelligent diagnosis.By studying the baseic principles,model srtructure and algorithms of artificial neural networks for a typical rotating machinery fault characteristics,choose the correct neural network model to do analysis and diagnosis can greatly improve fault diagnostic accuracy.