Microstructural Classification of Bainitic Subclasses in Low-Carbon Multi-Phase Steels Using Machine Learning Techniques

With its excellent property combinations and ability to specifically adjust tailor-made microstructures, steel is still the world’s most important engineering construction material. To fulfill ever-increasing demands tighter tolerances in today’s industry, research remains indispensable. The continuous material development leads more complex which especially true for designs that include bainitic structures. This poses new challenges classification quantification of these microstructures. Machine learning (ML) based microstructure offers exciting potentials this context. paper concerned with automated, objective, reproducible carbon-rich second phase objects multi-phase steels by using machine techniques. For successful applications ML-based classifications, a holistic approach combining computer science expertise domain knowledge necessary. Seven classes are considered: pearlite, martensite, subclasses degenerate debris cementite, incomplete transformation product, upper lower bainite, can all be present simultaneously one micrograph. Based on SEM images, textural features (Haralick parameters local binary pattern) morphological calculated classified support vector machine. Of objects, 82.9% correctly. Regarding total area 89.2% reported basis an improved, sophisticated quantification, enabling process–microstructure–property correlations established thereby forming backbone further, microstructure-centered development.