Research on the Modification of the Tapered Roller of Rotating Liner Hanger Bearing Based on Hybrid Intelligence

The heavy load causes the roller end breakage of the rotating liner hanger bearing, which is one of the important reasons of bearing failure. This paper researches on the internal shape modification of the tapered roller of bearing to solve the edge effect and extend the bearing service life. Firstly, a Back Propagation (BP) neural network was established where inputs are the degrees of depth and hollow and outputs are the maximum contact stress between the roller and the shaft ring of the bearing under different depth and hollow and the maximum equivalent stress of the roller. Secondly, the genetic algorithm was applied to optimize the BP neural network structure to prevent it from falling into the local minimum and obtained the best modification parameters. At last, finite element method was applied to simulate the predicted result. The simulation results showed that edge effect disappeared and the maximum contact stress between the roller and the shaft ring and the maximum equivalent stress of the roller were greatly improved.