Abstract: The kernel density estimation method is successfully used to solve the structural dynamic model validation challenge problem to further clarify the implementation process of model validation in structural dynamics. Probability distribution properties of sub-structure due to the weakly nonlinear connections and individual differences is not normal distribution in structural dynamic model validation challenge problem presented by Sandia National Laboratories, so kernel density estimation method is used to establish the sub-structure probability model and dimension reduction utilizing kernel principal component analysis is used to improve the computational efficiency. The model accuracy is qualitative verified and quantitative confirmed using calibration and validation experimental data respectively based on the probability sub-structure model. Finally the accreditation structure assessment and the target structure prediction are performed using validated substructure model, which are consistent with the results of other researchers. The research indicates that the kernel density estimation is an effective approach to solve the model validation problem in structural dynamics.

Key words: Challenge problem, Kernel density estimation, Kernel principal comment analysis, Model validation, Structural dynamics, Uncertainty analysis