Optimisation design of JM 3 wheel profile based on NURBS-NSGA-II algorithm for wheel-rail matching robustness

To address low-frequency carbody hunting of CR200J power-concentrated EMUs under specific track conditions, a multi-objective optimization design for the JM3 wheel tread is presented. Track monitoring tests and dynamic simulations reveal that the instability originates from the excessively low equivalent conicity of the JM3 wheel when matched with CN60N rails under large railcant. A wheel profile parameterization method based on Non-Uniform Rational B-Splines (NURBS) is proposed, in which the vertical coordinates of control points in the wheel-rail constant contact area are taken as design variables. The NSGA-II algorithm is employed to optimize the profile with respect to equivalent conicity concentration, wheel-rail contact stress, and contact point distribution, yielding an optimized profile denoted as JM3N. Results show that JM3N significantly improves the equivalent conicity of wheel-rail matching and reduces its sensitivity to variations in rail profile and rail cant. Under a 1/20 rail cant condition, the nonlinear critical speed increases by more than 100% compared with the original JM3 profile. Low-frequency carbody swaying is effectively suppressed, and both lateral ride comfort and curving performance are improved. Wear prediction further indicates that JM3N maintains superior dynamic performance after long-term operation, confirming the effectiveness of the proposed optimization.