Microstructure quality control of steels using deep learning

In quality control, microstructures are investigated rigorously to ensure structural integrity, exclude the presence of critical volume defects, and validate formation target microstructure. For quenched, hierarchically-structured steels, morphology bainitic martensitic major concern guarantee reliability material under service conditions. Therefore, industries conduct small sample-size inspections materials cross-sections through metallographers needle such microstructures. We demonstrate round-robin test results revealing that this visual grading is afflicted by pronounced subjectivity despite thorough training personnel. Instead, we propose a deep learning image classification approach distinguishes steels based on their microstructure type classifies length alluding ISO 643 grain size assessment standard. This facilitates reliable, objective, automated hierarchically structured steels. Specifically, an accuracy 96% roughly 91% attained for distinction martensite/bainite subtypes length, respectively. achieved dataset contains significant variance labeling noise as it acquired over more than 10 years from multiple plants, alloys, etchant applications, light optical microscopes many (raters). Interpretability analysis gives insights into decision-making these models allows estimating generalization capability.