Abstract: Wheel profile optimization is a good solution to the problem of reduced dynamic performance caused by wheel wear of high-speed trains. A rotary-scaling fine-tuning method(RSFT) is used to generate the new wheel profile; a vehicle dynamics model for the certain high-speed trains is established, and the corresponding optimization objectives and constraints are calculated by the model; the optimal profile is optimized using a radial-based neural network-particle swarm optimization(RBF-PSO) algorithm. By comparing the dynamics and wear performance of the wheel profile before and after the optimization, it can be found that:the wheel profile critical speed after optimization is 424.6 km/h, an increase of 10.2%; the lateral and vertical ride indexes are reduced overall, while the safety indexes during curve passing are improved, and the derailment coefficient, overturning coefficient and lateral axle force are further reduced. The optimized wheel profile has a more evenly distribution of contact points and a reduced equivalent conicity. At the same time, the optimized wheel profile effectively reduces the depth of wheel wear and reduces the root wear of the wheel rim, reducing the maximum depth of wheel wear by 9.8%.

Key words: high-speed trains, wheel profile optimization, multi-objective, wheel wear, wheel-rail contact