Global sensitivity evolution of vertical ride quality for high-speed maglev trains

This study investigates the influence of structural parameter randomness on the ride quality of high-speed maglev train levitation systems, aiming to provide a theoretical foundation for structural optimization and control strategy design.The complex levitation system is first simplified into a single electromagnet levitation system incorporating secondary suspension, and the Sperling index is employed to assess ride quality quantitatively.Based on the probability density evolution method (PDEM), the probability density function (PDF) of the Sperling index is estimated.Subsequently, global sensitivity analysis is conducted using the Frchet derivative, and the results are compared with those obtained using the conventional Sobol's method.The findings show that PDEM achieves a mean absolute error of only 0.0419 compared to the Monte Carlo method (MCM) with an 806fold improvement in computational efficiency.The sensitivity rankings obtained from the method based on Frchet derivative are consistent with those from the Sobol's method and further reveal the directional impact of parameters on the Sperling index.Overall, the electromagnetic winding turn, the magnet area, and the quality of the levitation electromagnet are identified as the key structural parameters influencing ride quality.