Evaluating the prediction accuracy of polynomial regression models for clinch joint properties
In modern lightweight design, mechanical joining processes such as clinching are becoming increasingly common. However, variations in tooling, process parameters, and material properties can cause significant deviations in clinch joint properties. In a joint study, the Chair of Engineering Design (KTmfk) at Friedrich-Alexander University Erlangen-Nürnberg and the Chair of Forming and Machining Technology at Dresden University of Technology present a two-stage evaluation of global polynomial regression metamodels. The models are first benchmarked against deviation simulations and then against experimentally observed scatter. While the metamodels accurately reproduce the numerical variation, the comparison with experiments reveals property-dependent discrepancies, which are attributed to accumulated uncertainties in the metamodels, the numerical simulations, and the experiments.
https://doi.org/10.1016/j.jmapro.2025.09.059