Corrosion Prediction Approach Combining FEM‐Simulations and AI‐Modules for Complex Automotive Geometries

ABSTRACT The necessity of corrosion protection significantly influences the development of automotive body‐in‐white assemblies. This paper presents a digital hybrid framework that enables early‐stage prediction of corrosion behavior in complex automotive geometries by combining physics‐based simulations with data‐driven artificial intelligence (AI) models. Validation on real vehicle components demonstrates high predictive accuracy, with correlation coefficients up to and the majority of results within a 30% deviation from experimental measurements. The approach thus provides a scalable and physically interpretable method for virtual corrosion evaluation, enabling data‐driven optimization of corrosion protection concepts in automotive design.