According the characteristics of rolling bearing fault information are nonlinear and non-stationary, the method of time-frequency is often used to make the one dimensional time signal map into two dimensional time and frequency function, and describe the energy density of signal at different times and frequency simultaneously. A method of fault diagnosis based on S transformation and image Hu of invariant moments was put forward in this paper. First of all the measured rolling bearing signals have been S transformed, and time-frequency spectrum which is got is expressed as two dimensional image, then Hu geometric moment invariant of the S transformation spectrum is calculated and the simulation research is carried out using invariant moment principle in image processing. The results show that this method can distinguish the inner ring, outer ring and bearing roller fault intuitively and accurately and measure rolling bearing fault diagnosis efficiently.

Electrolytic aluminium pre-baked anode conductive device for the special friction Welder upsetting system is the research object. Establish dynamic equation of the upsetting system by its Lagrange equation; analyze the key factors affecting the systems dynamic characteristics; build the systems finite element model to be real by ANSYS Workbench; Set key parameters for the two sets of plans and do analysis for the modal and harmonic response successively. The two plans compared results reveals that the guide spacing and the damping among all the parts can affect the systems dynamic characteristics, which will provide a basis for further improvement and optimization of special friction welding machine bed.