A novel stress manipulation method for robust and reliability-based optimization of multi-step forming processes

In this study, a novel framework is developed that treats local stressregion parameters (size, location, type, and timing) as explicit design variables for robust, reliabilitybased optimization of multistep metalforming processes. Addressing the gap that current workflows either rely on global process controls or manual adjustments and are computationally prohibitive for systematic robustness design, our approach couples an automated FEM-DoE pipeline with surrogate metamodels and screening techniques to reduce the number of full finiteelement evaluations. The workflow integrates meshsensitivity analysis, automatic multistep simulation and output extraction, a stressmanipulation algorithm, and downstream Pareto, stability, and spread analyses, and supports both robust and reliabilitybased optimization to select stressmanipulation size, location, type, and flattening depth according to user objectives. In a representative multistep forming case study, targeted stress manipulations combined with optimization reduced springback variability and improved robustness by up to about 60% relative to the unmodified baseline. Permitting flattening depth as an unconstrained design variable produced robustness gains up to around 80% in unconstrained scenarios, though with measurable reductions in reliability in those cases. Fullthickness, pretools opening interventions consistently outperformed partialthickness or posttools opening adjustments. These results demonstrate that local stress manipulation is an effective lever for improving robustness and that the insights gained from this research based on the proposed automated pipeline enables reproducible, computationally tractable, reliabilityaware decision making for industrial metalforming design where controlling parameters is critical to ensuring the accuracy and robustness of products.