Substructuring-based parametric reduced-order modelling for structural dynamic predictions of bolted assemblies.

Bolted joints play a crucial role in mechanical systems by connecting different components and materials. The performance of these systems heavily relies on the behavior of the bolted connections. High-fidelity finite element (FE) models are commonly used to predict their dynamic behavior, but they suffer from long computation times due to large model sizes and iterative solution procedures for contact nonlinearity. To address this, a substructuring-based parametric model order reduction strategy is proposed to improve the computational efficiency of FE models of bolted structures with frictional contact. The strategy involves dividing the model into substructures, including joint substructures, and applying a two-layer Krylov-based Craig-Bampton (CB) method for reduction. The augmented characteristic constraint modes, combined with joint interface modes, are proposed for contact interface reduction. This approach is valid for different numerical representations of contact and does not rely on full model results for bolted joints. Consequently, a sampling-free parametric reduced order model can be generated to analyze changes in pretension and friction coefficient. For time domain simulations, a dual loop contact algorithm integrated in the Newmark-Beta method is present for time integration of the parametric reduced model. The performance of the approach is validated through investigation of overall dynamic responses and local contact variations in two bolted joint structures. [ABSTRACT FROM AUTHOR]