Edge texture extraction method based on mathematical morphology of vibration parameter image for rotating machinery

The fact of a scarce consideration of multi-dimensions image information affects the diagnosis technology promotion and utilization in a certain extent. The method, through which the texture features in vibration parameter image for a rotating machinery are directly extracted and mined, is investigated in this paper. And then, a method based on Mathematical morphology is also presented. Mathematical morphology filter of open-close and close-open combination to realize vibration parameter image filtering is introduced. And different morphology edge detection operator is studied. The application effect of classical LOG edge detection operator and edge texture feature extraction method based on morphology gradient for the rotating machinery vibration parameter image are discussed. On the modeling of 600 MW turbine rotor experimental bench, rotor's normal state, fault of unbalance, misalignment, steam exciting vibration and bearing pedestal looseness are being examined. The diagnosis result shows this method can accurately extract obtain image feature. This is a new way for investigating rotating machinery fault diagnosis.