MaterialsWeek 2025
Lecture
03.04.2025
AI-Driven Material Design: Accelerating Innovation with SLAMD
SM

Sarah Munsch (M.Sc.)

Bundesanstalt für Materialforschung und -prüfung (BAM)

Munsch, S. (Speaker)¹; Klewe, T.¹; Kruschwitz, S.¹; Moreno Torres, B.²; Schumacher, K.³; Yared, K.³; Zia, G.A.J.¹
¹Bundesanstalt für Materialforschung und -prüfung, Berlin; ²Technische Universität Berlin; ³Holcim (Deutschland) GmbH, Sehnde-Höver
Vorschau
17 Min. Untertitel (CC)

In the Horizon Europe-funded project REINCARNATE, a user-friendly web application, the "Sequential Learning App for Materials Discovery" (SLAMD), was developed to revolutionize the materials design process. By integrating a digital lab twin with AI-driven approaches, SLAMD supports and significantly accelerates the discovery of optimal material compositions within complex parameter spaces, while considering eco-efficiency and cost-effectiveness. Its versatility and practical relevance are demonstrated through diverse applications:

SLAMD is being used to develop two classes of concrete incorporating recycled materials, each with unique requirements and constraints imposed by regulations and industry standards. The first application focuses on optimizing compressive strength with at least 30 % substitution of the CEM II binder using a mix of recycled glass powder and concrete fines. The second targets the formulation of concrete for five exposure classes, exceeding the regulatory limit of 50 % recycled aggregates while meeting all performance criteria. Besides finding new concrete mixes with a higher percentage of recycling materials in their composition, these demonstration cases aim also to spread the use of SLAMD as a tool for industry stakeholders.

In collaboration with the industrial partner Holcim (Germany) GmbH, SLAMD accelerated the development process of concrete with high carbonation resistance by a factor of 80 %. This success was further enhanced by incorporating NMR features from non-carbonated sister samples at only 28 days of age, revealing additional acceleration potential. Currently, the integration of non-destructive testing methods, such as Laser-Induced Breakdown Spectroscopy, is being explored to further shorten development cycles.

Another example is the use of SLAMD within the Volkswagen Foundation-funded project "Circular B-IO," where it is adapted and applied to optimize the heterogeneous material flow of recycled substances. By creating homogenized supply streams, SLAMD contributes to the reliable production of low-carbon concrete in Kenya, enabling more sustainable construction practices.

Looking ahead, SLAMD may be used to address a wide range of material-related challenges. Through its ability to navigate complex datasets and identify promising solutions efficiently, it opens new opportunities for advancing material design, accelerating innovation, and promoting sustainability.


Abstract

Abstract

Erwerben Sie einen Zugang, um dieses Dokument anzusehen.

Ähnliche Beiträge

© 2025