Research On Blind Source Separation Based On High-order Cumulant And It's Application On Rotating Machinery

The most important step of mechanical fault diagnosis is to extract the fault feature,Blind source separation has a bright scene in the field of fault diagnosis, and it is a multi-channel signal processing method, source signal can be found in the case of unknowing transmission channels.For mechanical fault diagnosis, the most important thing is extracting the information that can reflect the state of mechanical operation. Characteristic signal can be extracted in the operation of the various BBS algorithm .this paper will explore how the blind source separation applied to the mechanical to fault diagnosis.The main work is as follows:An introduction to blind source separation, blind source separation theory of the development and application in the development of fault diagnosis.Introduced the system involved in blind source separation theory to a variety of basic theoretical knowledge and a variety of blind source separation algorithm.About mechanical vibration signals in machinery and equipment in the transmission mechanism and a variety of common gear and bearing failures and frequency domain performance characteristics.The source of the estimate is a blind source separation is a very important preparatory work to do which describes a variety of signal estimation method, proposed a new signal processing methods.Based on second order cumulant cumulants joint diagonalization for blind source separation algorithm is a very stable algorithm for blind source separation, this paper describes the principles of the two algorithms and separation performance, experimental design, according to wilderness Separation model model, low-pass filter, the signal processing algorithm is then used in the separation of the two real signal acquisition.