Multi-variable control with feedforward control of product properties generated by forming processes - basics and application to stamp hole rolling

• Applicants: Professor Dr.-Ing. Peter Groche; Professor Dr.-Ing. Volker Schulze
• Subject Area: Primary Shaping and Reshaping Technology, Additive Manufacturing ; Mechanical Properties of Metallic Materials and their Microstructural Origins
• Term: since 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 424334856

Within the scope of this project, the previous fundamentals for the control of forming-generated product properties in punch-hole-rolling and the concepts for soft sensor measurement by means of magnetic Barkhausen noise (MBN) will be used to control the properties of the surface layer (especially hardness) and the product geometry in real time during the process. Here, findings from the investigations on the TRIP steel 1.4301 from the second funding period will be used to be able to adjust the surface hardness and the collar height by varying the speed and the radial feed during the process. The implementation of the controller on the already developed tool for servo motor presses is done by a cascade control, which uses a model predictive approach based on the data of a tool integrated measurement with the MBN soft sensor to initiate and monitor the controller parameters and to adjust the reference variables via the order reduced models found so far. The martensite content in the surface layer and the optically measured collar geometry are then used as actual values by the process controller, which uses the machine-provided degrees of freedom to respond to the command variables. In order to use the soft sensor as robustly and precisely as possible, other boundary layer state variables such as residual stresses and grain sizes must also be recorded in order to implement knowledge of their surface layer contributions comprehensive and as accurate as possible. In order to derive further material models to control these variables, the surface layer states in the parameter study must therefore be investigated and locally resolved by means of metallographic, X-ray and electron microscopic analyses. On the basis of the FE simulation, which will be supplemented and validated by material models in the second funding period, a prediction of the depth information of the microstructure state will be made possible in addition to the analysis of areas directly at the surface. This can then simultaneously support simulation-based control over the entire product properties. Subsequently, the disturbance influence of the semi-finished product variations on the product properties will be quantified. For this purpose, classical disturbances in the semi-finished product such as the microstructure condition, the sheet thickness and the starting hole diameter are to be mapped in their effect on the soft sensor signal in order to enable control even in the case of initial or discontinuous fluctuations of the controller variable. Finally, the controller quality will be evaluated on the basis of the optimization capability with respect to the wear properties as a function of the surface layer condition. Finally, generalized design principles for property-controlled forming processes will be derived.

• DFG Programme: Priority Programmes
• Subproject of: SPP 2183: Property-controlled metal forming processes
• Co-Investigator: Dr.-Ing. Stefan Dietrich