Abstract: In order to overcome the nonlinearity and instability of the shaft riveting assembly process of wheel hub bearing units,an optimization strategy based on the surrogate model was proposed for the optimization design of the rivet head.In the investigation,first,the finite-element modeling parameters were determined through the tensile test,the ring upsetting test,the on-site axial feeding displacement test,the on-site axial riveting force test and the theoretical derivation about the space trajectory of the rivet head.Next,a comparison between the simulated and the tested results of the axial riveting force and the hub shaft's ultimate deformed shape was made to verify the effectiveness of the proposed finite element model.Then,on the basis of the existing design experiences and the life theory of hub bearing units,the parametric equations and the optimization objective were determined.Moreover,on the basis of the Latin hypercube sampling strategy and the finite element simulation results,a support vector machine regression model,which takes into consideration the maximum radial deformation of inner ring,the axial preload,the acting force between the hub shaft and the rivet head as well as the forming surface parameters of the rivet head,was established.Finally,a verification of the optimization design was performed,with an improved performance of the hub bearing unit being achieved.

Key words: hub bearing unit, shaft riveting assembly, surrogate model, support vector machine regression