Feasibility study - Material Property Characterization Platform

Important results from the project

During the project, we developed a strategy for an AI-based material property prediction platform, tested different algorithms and were able to show that the prediction of the material property worked well for a limited group of steel grades and production processes. Limitations within the available data sets meant that we were able to clearly define the eligibility requirements that are set for a possible full-scale project. A project plan for a possible full-scale project has been developed.

Expected long term effects

Based on the project´s results, we can conclude that the material characterization process at the participating companies can be shortened in time and costs. In particular, being able to reduce the number of physical material samples for different alloys in different production states will have a positive effect in the long term. This will facilitate a faster introduction of new steel grades and alloys.

Approach and implementation

First, we defined the limitations of material, applications and material properties, where the choice fell on AHHS for crash-relevant vehicle components with a focus on yield stress, yield strength and elongation at fracture. Next, we focused on the availability and structure of existing material data from consortium members. Then, we conducted a literature study on algorithms for predicting material properties, tested and recommended some. We validated the results with material tests.