Paving the way towards digital materials selection using AI and physical models (DMS-AI)

Purpose and goal

The project aims to integrate AI and physical models for predicting complex material properties to improve materials selection. The project will explore and evaluate AI methods for scalable, effective material property prediction, leading to a framework that advances sustainable material selection across a wide range of applications. The pre-study aims to show how AI methods, existing data, and physical simulation, can be combined to predict properties like wear, fatigue, and corrosion.

Expected effects and result

The project anticipates delivering a detailed roadmap for AI-based material property prediction, including feasibility studies and practical recommendations for implementation. Outputs include a technical report reviewing existing methods, challenges, and opportunities, and critical factors needed to optimize data and model selection. The expected outcome is a foundation for a standardized approach that efficiently leverages AI for sustainable material selection and property prediction.

Planned approach and implementation

The project will begin with a survey of current data and AI techniques in materials prediction, and workshops to assess industry needs as well as different initiatives in the area. Next, it will review and map AI models suitable for selected properties, conducting feasibility studies on selected properties to evaluate data compatibility and accuracy. The final phase includes creating a consortium and a roadmap detailing steps for scalable AI integration.