Automated mechanical lab' and model calibration – AutoMeCal
Automated Mechanical Lab’ and Model Calibration
Automated mechanical lab' and model calibration
Aims of the AutoMeCal project
Extensive use of virtual mechanical design for shaping processes and structural parts makes it possible to use materials more sparingly, reducing product mass and production rejects due to variations in material properties from one batch to the next. The robustness and reliability of such a design is highly dependent on the accuracy of the mechanical model of the material, defined by constitutive equations and optimised parameters for the chosen material based on data from mechanical tests. An automated solution for mechanical testing and calibration of mechanical models makes it possible to study the impact of a variation in mechanical properties from one batch to another virtually, and therefore relatively quickly. The aim of the AutoMeCal project is to create an automated and intelligent tool, the input of which is a new material, and the output of which is the anisotropic elasto-plastic model calibrated to failure, which is precisely representative of the mechanical behaviour of the material. This tool includes a robotised cell for automated mechanical testing, coupled with a fully automated optimisation tool based on the inverse identification of material parameters. The innovative aspect of the AutoMeCal project is the integration of several components and/or tasks that, until now, already exist on their own but do not work together in a fully automated way.
The AutoMeCal project began on 1 February 2024 and will run for 4 years. The tool to be developed will comprise a robotised cell for automated mechanical testing, consisting of a simple tensile or shear testing machine and a cobot, coupled with a fully automated optimisation tool based on the inverse identification of material parameters. The aim of the AutoMeCal project is to integrate a number of components and/or tasks, such as specimen preparation and pattern deposition for digital image correlation, mechanical testing and validation, data preparation for identification and the choice of a method (Finite Element Model Updating or Virtual Fields Method) for calibration, which already exist on their own but need to be developed and/or adapted for automation. This tool is a strong contribution to Industry 4.0, as the quality of input data for numerical simulation is essential for the robustness and accuracy of numerical predictions. There is a recognised lack of databases on materials, and the AutoMeCal project will help to improve the quality of numerical simulation, particularly in the transport sector (automotive and aerospace) and for thin metal sheets.
The tool to be developed will comprise a robotized cell for automated mechanical testing, consisting of a simple tensile or shear testing machine and a cobot, coupled with a fully automated optimization tool based on the inverse identification of material parameters. The aim of the AutoMeCal project is to integrate a number of components and/or tasks, such as specimen preparation and pattern deposition for digital image correlation, mechanical test execution and validation, data preparation for identification and choice of method (Finite Element Model Updating or Virtual Fields Method) for calibration, which already exist in their own right, but need to be evolved and/or adapted for automation. This tool is a strong contribution to Industry 4.0, as the quality of numerical simulation input data is essential for the robustness and accuracy of numerical predictions. There is a recognized lack of material databases, and the AutoMeCal project will help to improve the quality of numerical simulation, particularly in the transport sector (automotive and aerospace) and for thin metal sheets.
/
Souto, N.; Andrade-Campos, A.; Thuillier, S. Material parameter identification within an integrated methodology considering anisotropy, hardening and rupture. Journal of Materials Processing Technology. 2015, 220, 157-172.
Kacem, A.; Laurent, H.; Thuillier, S. Influence of experimental boundary conditions on the calibration of a ductile fracture criterion. Engineering Fracture Mechanics. 2021, 248, 107686.
Hérault, D.; Thuillier, S.; Lee, S.Y.; Manach, P.Y.; Barlat, F. Calibration of a strain path change model for a dual phase steel. International Journal of Mechanical Sciences. 2021, 194, 106217.
Ouaidat, G.; Kacem, A.; Thuillier, S. Uncertainties on the mechanical behaviour of bronze sheets: influence on the failure in bending. International Journal of Material Forming. 2024, 17, 29.
In the field of virtual mechanical design, the robustness and reliability of a mechanical model depends strongly on the experimental database used for the calibration, I.e., the richness/completeness of the database and the closeness to the material used in production. At the moment, within an industrial context, the link between the virtual material and the real one used in production is not strong enough, and at the best, is obtained on other batches, in different aged states or thicknesses. The aim of the AutoMeCal project is to provide, rapidly and efficiently, a calibration of advanced mechanical models from the batch used in production. To this end, an automated mechanical characterization and model calibration will be developed, to provide, by an inverse methodology, the material parameters needed for the virtual design of the forming process and/or the structural design.
Project coordination
Sandrine THUILLIER (Institut de Recherche Dupuy de Lôme)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
Partnership
IRDL Institut de Recherche Dupuy de Lôme
Help of the ANR 292,177 euros
Beginning and duration of the scientific project:
January 2024
- 48 Months
