Dynamic analysis and correlation propagation of the multibody system considering correlated interval parameters

Interval uncertainty quantification for multibody systems has gained increasing attention due to complex requirements in the dynamic analysis of virtual prototypes. It is important to carefully consider the correlation between input interval parameters to avoid overestimating predictions, which can happen in traditional interval analysis that assumes parameters are mutually independent. This study introduced a new method that propagates multiple interval parameters with large uncertainty levels in multibody systems. The method combines local mean decomposition and bivariate function decomposition with Chebyshev polynomials to create a coupled surrogate model. This model can envelope the examined interval response and calculate response correlation coefficients. Numerical examples are provided to demonstrate the effectiveness of the method. The results showed that the proposed approach efficiently handles multiple interval parameters with significant uncertainty in the dynamic analysis of a multibody system.