Project description

The goal of the project is to develop a holistic modeling method for simulating the process chain for the production of complex bend aluminum profiles. Critical-to-quality parameters (CTQs), e.g. springback, wall thickness, energy balance of the process chain, and mechanical performance, as well as uncertainties will be modelled and predicted. This lays the foundation for optimizing the process parameters throughout the entire process chain in the second phase. 

The process chain includes extrusion, bending and heat treatment of aluminum profiles for automotive components. Global optimization of CTQs is difficult to achieve, as existing process models are not structured to be sequentially linked. Uncertainties stemming from process conditions, material properties, measurement errors, and limited prior knowledge, to list a few, further increase the complexity of the problem. The project follows a two-stage modeling strategy: In the first stage, physical white-box models are applied to gain a more comprehensive understanding of the process chain and interactions between process parameters. In the second stage, more efficient surrogate models will be built as substitutes for the physical models and experiments. 

The main objectives of the research project are: 1) Identification of critical stages or parameters in the production process, along with an assessment of their predictability; 2) Establishment of the process chain to investigate the relationships between process parameters and CTQs; 3) Building surrogate models for expensive numerical models and experiments. 4) Evaluation of uncertainties, with a particular focus on identifying and eliminating their reducible components. 5) Global optimization of both CTQs and the robustness of the process chain under non-reducible uncertainties, achieved by quantifying and integrating aleatory and epistemic uncertainties.

Process chain