Predicting Intergranular Corrosion Penetration Depth in Accelerated Testing using Machine Learning

Kurzfassung

Aluminum (Al) 6000 series alloys are relevant in automotive applications due to great strength, excellent machinability, formability, as well as good weldability. Nevertheless , aluminum alloys are susceptible to intergranular corrosion (IGC) making it crucial to determine the relationship that exists between the alloy composition and manufacturing parameters to the corresponding rate and magnitude of IGC observed. The large number of parameters that affect this behavior requires the development of a more robust framework. In this work, a probabilistic model is leverage to establish an accurate linkage between a set of input parameters (e.g., alloying content, heat treatment, etc.) and the resultant IGC performance (i.e., penetration depth) of multiple Al 6000 series alloys. The model extracts usable knowledge from a curated database comprised of 128 individual experimental results. The developed framework and established models are further validated by predicting the penetration depth of six new/unseen alloys.