Fault Diagnosis Of Gear-box Based On Wavelet Neural Network Expert Systems

Gearbox plays a very important role in connection and transmission power,and applys abroadly in mang mechanism.According to the statistic,the fault of gear plays about 10% among the whole fault in the mechanism,and the fault ratio of the bearing would possess a huge proportion.So,when the failure diagnosis of gear-box becomes a very important part of mechanism,the research of gear and bearing about failure diagnosis has a significance in the practice work.The importance of Gearbox-Fault-Diagnosing consists not only in cutting the time and cost on repairing,but also increase the stability of mechanical system and security of systems operation.Therefore,it is full of meaning.At present,the fault diagnosis has been developped to the intelligentized field.It is not only promote the intelligentized diagnosis methods,but also has been a new and very important means in the the research of fault diagnosis technology.In this paper, the packet de-noised theory is used to foreclose to the fault signal,at the same time to pick up the energy character successfully which is becoming the input parameter of neural network expert systems(NNES).It has been a new method in telling apart the fault modes of signals.The hide repository, which is trained the net of NNES by the training stylebook,is transfered to make the conclusion of fault diagnosis through the testing stylebook.Selecting five fault modes as the studing objects in the gearbox..The results of the experimentation show that the wavelet neural network expert systems has a better net performance and can apply to the fault diagnosis technology of the practice engineering.The intelligentized diagnosis program about gearbox is designed which combines MATLAB and VB,that has humanization interface and is easy to handle.Through the paper,it is not only show the mightiness function of MATLAB in calculating,ploting and neural network toolbox,but also exert superiority of VB in the user interface sufficiently.