An intelligent iterative framework for rivet-die structural multi-variable optimization in slug rivet aeronautical assembly

This study develops an intelligent, automated optimization framework that integrates parameterized finite element (FE) modeling, automated dataset construction, genetic algorithm (GA)-based multi-objective optimization, and machine learning approach was developed for the multi-variable rivet-die structural optimization for slug rivet assembly. By systematically exploring complex design spaces and automating data processing with tools such as modeFRONTIER, ABAQUS, and Python, the framework enables rapid, high-fidelity evaluation and optimization of structural parameters. Quantitative results demonstrate the effectiveness of the proposed methodology: the maximum interference level was elevated by 45.51% compared to the baseline model, increasing from 2.670% to 3.885%. The method also substantially improved the distribution uniformity. Experimental validation confirmed the accuracy of the simulations. These results indicate that the proposed intelligent framework not only delivers substantial improvements in interference quality and process efficiency, but also provides a scalable solution for data-driven structural optimization in advanced digital manufacturing contexts.