Linear vs. Nonlinear Bolt FE Modelling: a Methodological Approach for Improved Simulation Reliability in Automotive Design

Bolted joints represent one of the most widespread connection technologies in mechanical and automotive structures, where reliability and accuracy in their numerical modelling are crucial for predicting the global response of assemblies under different loading conditions. Despite their apparent simplicity, bolted joints exhibit a complex mechanical behaviour driven by the interaction between preload, contact mechanics, friction, and possible nonlinearities due to material plasticity and local separation phenomena. For this reason, the choice of an appropriate modelling strategy in finite element analysis (FEA) can strongly affect both the accuracy of the results and the computational cost of the simulation. This thesis investigates and compares linear and nonlinear finite element modelling approaches for bolted joints, with the objective of providing a methodological framework to assess their applicability, limitations, and accuracy in the context of automotive design. To this end, three case studies of increasing complexity are analysed: Case study I: Elementary single-lap joints, serving as benchmark configurations to evaluate the fundamental mechanical response of bolted connections. Case study II: Shock absorber top mount, a thin-walled steel component bolted to an aluminium flange representative of a critical load transfer element in suspension systems. Case study III: Vehicle body structure, consisting of the chassis and the front axle carrier of a passenger car, in which multiple bolted joints interact within a complex load path. For Case Studies I–II, high-fidelity nonlinear reference models are built in ABAQUS/Standard, explicitly accounting for bolt preload, evolving contact with friction, and geometric nonlinearity when relevant. These serve as benchmarks for simplified linear strategies implemented in Nastran. For Case Study III, given scale and cost, only linear models are considered. The linear strategies systematically evaluated are: • BMW approach, where bolts are represented by an idealized rigid connection; • RBE2+CBEAM strategy, where fasteners are simplified as a beam with the same nominal diameter of the bolt, coupled to plate nodes via rigid elements; • Stiffness-calibrated CBUSH elements, connector properties derived from analytical spring-analogy formulations and semi-empirical shear flexibility models. Comparisons are performed in the static domain via load–displacement and stiffness–force curves, and in the dynamic domain via frequency response functions: direct steady-state analysis in ABAQUS and SOL108 in Nastran, ensuring consistent excitation and damping assumptions. Results quantify the accuracy–cost trade-offs of each linear strategy against the nonlinear references and highlight the impact of modelling choices (e.g., bolt diameter representation, connector stiffness calibration, contact idealisations) on predicted joint stiffness and load transfer. Overall, the work provides a practical, evidence-based procedure to select bolt modelling strategies in automotive FEA, improving simulation reliability while controlling computational effort—particularly when migrating from detailed nonlinear models to reduced linear models in large assemblies and full-vehicle applications.