Environmental and economic constraints are pushing for a constant weight reduction of the exchangers, hence leading to a downgauging of the materials. Furthermore, emission regulations induce an increase of the operating conditions. Therefore, critical conditions are reached regarding the use of aluminum. It is important to be able to optimize the design of the exchangers with a simulation tool in order to design quickly and to avoid lengthy and expensive tests or prototypes. The failure mode of the exchanger results from thermal shock which is induced by transient differential thermal dilation of the exchanger components. This alternate strain results in low cycle fatigue (a few thousand cycles) on low gauge components. Furthermore, the manufacturing process of the exchanger (brazing) and the process of the aluminum low-gauge sheet has an important impact on the microstructure of the aluminum alloys. It is necessary to take into account the grain or sub-grain structure of aluminum in order to define the fatigue damage mechanism and the mechanical properties. The first step of the project is to define precisely the link between the manufacturing processes of the exchanger and the aluminum sheet, and the composition and microstructure of the aluminum alloy. Simultaneously, experimental measurement and improvement of the global exchanger simulation will allow determining the loading of the exchanger. A test bench will be setup in order to measure locally and with high precision the stress and mechanical properties. It will also allow creating samples with controlled damage for further analysis. The link between the loading conditions and the mechanical properties of the aluminum alloys will allow to define thermo-mechanical and metallurgical behavior laws. These laws will also include the non homogeneity of the assembled structure and the microstructure properties. The damage analyses will lead to setup low cycle fatigue criteria for low gauge brazed aluminum sheets. These criteria will be included into the global exchanger durability simulation. At the end of the project, the knowledge on the link between the microstructure and the fatigue resistance will allow the exchanger to reach better fatigue performance. Furthermore, virtual prototyping will be realized for industrial application taking into account precise loading conditions, low cycle fatigue damage criteria, manufacturing history and microstructure impact. Owing to the simulation tool, production time will be reduced

RECYCAL project aims at understanding and modeling the damage mechanisms operating in recycled aluminium alloys when loaded in plane strain mode. In comparison with reference alloys currently used in transport and packaging sectors, these alloys have a higher content in impurities, especially iron, which often leads to decreasing mechanical performance. RECYCAL is a partnership between Constellium company, Center of Materials, CEMEF and SIMaP laboratory. Its main objective is to predict the impact of microstructural features related to intermetallic particles (size, morphology, nature, spatial distribution) and their environment (matrix’ hardness, crystallographic orientations of grains) on ductility by combining in-situ 3D characterizations at different resolutions and multi-scale simulations. The effect of loading path will also be investigated in-situ by comparing plane strain tension and bending on miniaturized samples under X-ray synchrotron beam. Extension to complex industrial loadings, which are combination of these elementary loadings, will be performed by finite element simulations. The project will be organized in six complementary work packages: 1- Materials fabrication and initial characterization; 2- Observation and quantification of damage in plane strain tension; 3- Observation and quantification of damage in bending; 4- Full-field micromechanical simulations from tomography images of real microstructures; 5- Building of a damage criterion including relevant microstructural features coming from both in-situ observations (WP2 and WP3) and simulations on representative volume elements (WP4) (comparison between a standard statistical approach and a neural network); 6- Upscaling to industrial cases thanks to finite element simulations using the criterion proposed in WP5.