AZTERLAN and VEIGALAN develop optimized foundry materials by means of Artificial Intelligence
This advanced development has been achieved within the DigiMAT Project, aimed at manufacturing security brake system castings for passenger vehicles in close collaboration with the companies AAPICO and CONTINENTAL TEVES.
Gaining control over the manufacturing processes is necessary to produce high quality safety components that fulfil as well, the required properties. Process data analysis and Models Predictive Control (MPC) are some useful tools that allow to relate specific real time manufacturing parameters and the corresponding results, ensuring that process variables keep within the optimal operational limits, in order to achieve specific characteristics in the final components.
By applying these technologies to cast iron process, AZTERLAN Technology Centre and VEIGALAN technology-based company, along with the automotive sector companies AAPICO (cast iron foundry) and CONTINENTAL TEVES (Brake system TIER1 company), are working on an intelligent solution for producing materials with enhanced properties for the automotive industry. The technology developed up to date by the working team has allowed to optimize the metal alloy used to produce braking systems with the aim of manufacturing more resistant and lighter components for passenger vehicles.
In the words of Jon Garay, AZTERLAN researcher specialized in foundry technologies, “iron casting is a complex transformation process where a large number of variables are involved. Due to the excellent mechanical properties, the quality-cost ratio and the possibility to obtain complex geometries and near net shape parts, cast iron plays and will keep on playing a determining role in cars. Nevertheless, the current weight-lightening trends and requirements are pushing actual materials to new limits. To achieve this core objective, we need to be able to settle more robust and efficient processes”.
In a first phase, the working team has developed an artificial intelligence system that in real time connects manufacturing process data with the results obtained in the testing benches of AAPICO and the advanced characterisation studies performed at AZTERLAN. Thanks to the models developed using SALOMON artificial intelligence system, acting over specific parameters identified as relevant, “such as, the iron metallurgical quality, the chemical composition and other critical parameters of the process” the team has improved the properties of the material by improving its yield strength and maximum load. This is allowing to reduce the weight of the brake components.
As explained by the researcher of VEIGALAN Asier González, “we have focused our efforts towards structuring a digital architecture composed by a network of sensors, destructive and non-destructive testing controls, MPC and supervised deep learning algorithms to manufacture metal components with optimised characteristics”.
The technology that is being created within the DigiMAT project can be also applicable to other foundry materials and it has a direct impact on the design possibilities of complex components, promoting weight reduction strategies. Also, thanks to being able to assure that parts meet specific requirements during the manufacturing process, this new development makes it possible to eliminate specific post-production controls and non-added value operations.
The project consortium considers that this innovation: “DIGITAL MATERIALS” is potentially applicable to any other metals, opening opportunities to a new generation of optimized components, as well as to building more sustainable and efficient processes.
Currently, the project team is working on the implementation of the new models in AAPICO to afterwards, start manufacturing the first prototypes based on them.
DigiMAT project is funded by the EIC Fast Track to Innovation Horizon 2020 program of under the Grant Agreement No 830903.