Experimental and numerical investigation of adhesively bonded joint for high volumes automotive applications
Abstract
The automotive industry's increasing adoption of lightweight materials like carbon fiber composites, especially employing Sheet Molding Compound for high productivity, necessitate advanced joining technologies for high structural performance and safety. This thesis presents an integrated methodology combining experimental characterization, numerical modeling and optimization to develop a predictive model for structural adhesives, specifically addressing manufacturing-induced non-uniformities in bondline thickness. The research focuses on the quasi-static and fracture behavior of bonded joints, primarily using ductile polyurethane-based adhesive commonly employed with carbon fiber components. An extensive experimental campaign, utilizing standardized tests under various conditions (temperature, bondline thickness, surface preparation), provided data for calibrating and validating Finite Element models in Abaqus. Cohesive Zone Models were implemented to simulate fracture initiation and propagation. An inverse calibration procedure, facilitated by HyperStudy optimization software, was developed to systematically generate a validated "material card" (a set of constitutive parameters) that accurately describes adhesive behavior in Finite Element simulations, relevant to large-scale automotive models. This work addresses the lack in the literature of adhesive models for bonded SMC joints that are both predictive and transferable to industrial-scale FE simulations by explicitly considering manufacturing-induced bondline thickness variability and by developing an optimization-based cohesive zone modeling approach calibrated and validated using multiple experimental test configurations.
