For future sustainable high-performance materials, a transition is needed from stand-alone simulation and characterization tools towards integration of databases, physical modeling, inverse design, machine learning and experimental testing in a common framework accessible to all contributors and stakeholders. Such platforms will meet the challenges related to the development and optimization of materials and processes in high-dimensional design spaces and significantly accelerate identification of new materials.

ALPmat is an Active Learning Platform for materials design, aiming at connecting academia and industry and integrating all necessary digital tools for design and optimization of future materials. For optimal knowledge exploitation, hybrid approaches are employed, where physical models and expert knowledge are combined with data from observations and AI methods. The resulting hybrid models are used for Active Learning Loops (ALLs) to improve the addressed properties in an iterative way via optimization of the material’s chemistry and processing conditions, while also minimizing the number of new samples ensuring sample efficiency. We will present the details of ALPmat in terms of hard- and software for the platform backbone, the FAIR database, the framework for running physical modeling and Bayesian optimization algorithms, and integrated software services. Moreover, new data models tailored for our use cases are presented, linking all steps from synthesis and processing to characterization with respective metadata and capturing the digitized sample history. Finally, the application of ALPmat for identification of high-performance bainitic steels is shown. We performed a multi-objective optimization of the uniform elongation and the yield strength as a function of chemical composition and processing conditions. Our Pareto front outperforms state-of-the-art bainitic steels.