Research On The Neural Network And The Grey Theory In The Fault Diagnosis Of The Transfer-box

The transfer-box is the mechanical equipment in common use, which is used to change rotate speed and transfer dynamics, the aim for the condition testing and fault diagnosis of its transmission system is to offer scientific foundation for the pertinence maintenance, and save maintenance costs. In the fault diagnosis of the transfer-box, to pick up the fault information effectively and judge the fault property and origin properly, selecting the valid signal process and fault diagnosis method is the key to the success or failure of the diagnosis, the improving diagnosis sensitivity. In this dissertation, based on the analysis of the traditional diagnosis method, the facet of the intelligent diagnosis of the transfer-box is explored valuably, the technology of the artificial neural networks and gray system theory is applied to the fault diagnosis of the transfer-box. In this dissertation, based on the analysis of present research situation, the fault mechanism and diagnosis method of the transfer-box in the practical state of work are studied. The type of the transfer-box is JZQ250 and made as test objects. Firstly, based on the fault mechanism and vibration trait of the transfer-box in the practical state of work, the experiment project on condition testing and fault diagnosis of the transfer-box is designed, the vibration signals of the transfer-box are tested. At the same time, according to the traits of the vibration signals containing lots of noise, the testing signals are dealt with by the use of the noise cancellation in wavelet. Then after the noise cancellation in wavelet, the susceptibility of characteristic parameters in the vibration signals of the transfer-box are extracted, which contain the running characteristic information, the intelligent fault diagnosis of the transfer-box is studied, the system of transfer-box's state check and fault diagnosis through the improving BP artificial and ABO gray cognate analysis is established. From the result of simulation and experience in the dissertation, we can see the intelligent diagnosis method, which is presented by this dissertation, is effective.